Preparation of Highly Functional Spheroid of Endocrine Cells Based on Thermosensitive Glycol Chitosan

Abstract

Background:

Pancreatic islet transplantation holds great potential as a therapeutic approach for treating type 1 diabetes mellitus (T1D). However, large islets suffer from hypoxia due to the limited diffusion distance of oxygen, leading to cell loss. Therefore, smaller spheroids are needed for better transplantation outcomes. This study aims to develop a method for forming highly functional islet spheroids using glycol chitosan (GC) derivatives, such as N-acetylated glycol chitosan (AGC) and N-hexanoyl glycol chitosan (HGC).

Methods:

Thermogelling polymers were produced by performing N-acylation of GC using the correspondingly carboxylic anhydrides. Islet spheroids were formed using a dual application with AGC-coated plates and HGC gelation. The AGC solution was applied to the plate for coating and evenly distributed using a 1 mL syringe. Then, the HGC encapsulated with islet single cells was cultured on top of it. Spheroid viability and functionality were evaluated using CCK-8 assay and glucose-stimulated insulin secretion assay.

Results:

The aqueous solutions of AGC (4%, w/v) and HGC (36% hexanoylation) (2%, w/v) demonstrated a sol–gel transition temperature around 37 °C, suitable for the physiological environment. These polymers also showed no cytotoxicity to intact islets. Islet single cells were cultured on HGC gels with varying degrees of hexanoylation (DH) values, where higher DH values led to smaller and more uniform spheroids. The resulting spheroids formed on AGC-coated plates and HGC36 gelation were smaller and more uniform than those formed on untreated plates. These spheroids exhibited significantly improved glucose responsiveness, with superior insulin secretion.

Conclusion:

The optimized method using AGC and HGC offers a more efficient way to produce smaller, uniform, and functional spheroids.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13770-025-00708-x.

Introduction

Pancreatic islet transplantation is a clinically promising therapy for the treatment of type 1 diabetes mellitus (T1D). However, the viability of isolated islets from the pancreas, typically large (150–300 μm), is compromised due to hypoxia caused by the restricted distance of oxygen diffusion, leading to an immediate loss of approximately 70% of the transplanted islet mass. As a result, smaller islets (70–150 μm) may be more advantageous than larger ones in terms of viability under hypoxia [1, 2]. Dissociating the islets into single cells and reaggregating them into smaller spheroids allows for the formation of uniformly sized cell clusters, improving oxygen and nutrient diffusion, which in turn enhances their functional performance post-transplantation. Nevertheless, a major limitation remains as the re-aggregated islet spheroids fail to completely restore their functionality, particularly insulin secretion in response to varying glucose concentrations [3]. This issue arises from the random and broad size distribution of the spheroids, which occurs due to the absence of a driving force in a conventional flat culture dish [4]. Therefore, a novel strategy is needed to produce uniform, smaller, and functionally effective spheroids.

Conventional spheroid formation methods, such as micro-molding, hanging drop, and multi-well hanging drop plates, and centrifugation, have generated in vivo-like 3D cellular environments that mimic physiological tissues [5–7]. However, these systems frequently demand specialized microdevices, limit large-scale production, disrupt extended culture periods, and involve complex processes to retrieve spheroids for further analysis [8]. The ideal method for spheroid production should be efficient, scalable for large-scale manufacturing, and support long-term cell cultivation while allowing for easy spheroid retrieval. To date, non-adherent polymers have been employed to create and maintain 3D spheroids [9, 10]. In this approach, cells avoid adhering to the substrate and instead form spheroids on the non-adherent surface, where cell–cell interactions dominate over cell-substrate interactions.

Chitosan, derived from the deacetylation of chitin, has mainly been employed as a non-adhesive polymer to promote the generation of self-assembled 3D cellular spheroids [11, 12]. However, a key limitation of using chitosan as a biomaterial is its poor solubility under physiological conditions [13]. Glycol chitosan (GC) is a chitosan derivative easily soluble in aqueous solutions at any pH level and has free amine groups that can be modified chemically [14, 15]. Furthermore, its low toxicity and excellent biocompatibility make it ideal for use in biomedical applications [16, 17].

Specifically, our group has reported new classes of GC prepared by N-acylation, including N-acylated glycol chitosan (AGC) and N-hexanoyl glycol chitosan (HGC), which exhibit unique properties suitable for spheroid formation [18]. AGC is produced through N-acetylation of GC, introducing acetyl groups (–COCH₃) that enhance its non-adherent properties. Similarly, HGC is synthesized via N-hexanoylation of GC, incorporating hexanoyl groups (C₆H₁₁CO–). These modifications provide non-adherent and thermosensitive polymers designed to facilitate the formation of high-functionality islet spheroids. The non-adherent properties of AGC and HGC minimize cell-substrate interactions, allowing cells to aggregate naturally into spheroids. By modulating hydrophobicity, these polymers create an optimal environment for promoting cell–cell interactions, ensuring the formation of smaller, uniform spheroids with enhanced functionality. Their thermosensitive characteristics enable a controlled sol–gel transition at physiological temperatures, simplifying spheroid retrieval without mechanical disruption or enzymatic treatment, such as trypsinization, which preserves cell integrity [18]. In contrast, conventional methods like the hanging drop technique or micromolding often involve labor-intensive processes or require enzymatic and mechanical interventions, leading to cell loss, mechanical stress, and reduced viability [19–22]. Therefore, we anticipate these materials enable the retrieval of islet cell spheroids with high yield and viability through simple temperature regulation [23]. Consequently, we hypothesized that this approach could address the limitations of conventional methods while offering a scalable and efficient platform for islet cell spheroid production.

In this study, we synthesized the AGC and HGCs according to the degree of substitution of GC’s hexanoylation. Next, we evaluated spheroid’s formation methods (coating / gel / coating + gel) using AGC or HGCs related to uniformity, size, and functionality. In conclusion, we developed highly functional 3D islet cell spheroids using chitosan derivatives with non-adhesive and temperature-sensitive properties, enabling straightforward and efficient spheroid formation (Fig. 1).

Fig. 1.

Cell spheroid formation using AGC and HGC with non-adherent and thermosensitive properties. A Schematic illustration of the spheroid formation process on an AGC-coated plate with HGC gel. B, C Comparison of changes in the morphology of single cells. The single cells on the AGC-coated plate with HGC gel before (B) and after thermogelation (C). After thermogelation, the AGC prevents early cell adhesion to the plate and allows for controlled aggregation. HGC tightens cell–cell interactions, promoting spheroid formation

Materials and methods

Material and reagents

GC (DP ≥ 200; degree of acetylation = 9.3%) was obtained from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Propionic anhydride (99%), butyric anhydride (98%), Acetic anhydride (99.5%), valeric anhydride (97%), hexanoic anhydride (97%), medium 199, histopaque-1077, fetal bovine serum (FBS), and trypsin–EDTA were purchased from Sigma-Aldrich (Saint Louis, MO, USA). Methanol and acetone were supplied by Daejung Chemical Co. Ltd. (Gyeonggi, Korea).

Dialysis membrane (Molecular weight cut-off: 10–12 kDa) was obtained from Spectrum Laboratories (Houston, TX, USA). All chemical reagents were of analytical grade and used without additional purification. RPMI-1640 medium, LIVE/DEAD™ Viability/Cytotoxicity Kit, and Pico Green Quantitative DNA Assay Kit were purchased from Invitrogen (La Jolla, CA, USA). Penicillin–streptomycin was obtained from Gibco (Carlsbad, CA, USA), and the Cell Counting Kit-8 (CCK-8) was sourced from Dojindo Molecular Technologies (Rockville, MD, USA).

Synthesis of AGCs and HGCs

N-acetyl glycol chitosan (AGC) and N-hexanoyl glycol chitosan (HGC) were synthesized following a previously described method [18]. In brief, they were produced by N-acylating glycol chitosan (GC) with the corresponding carboxylic anhydrides. To synthesize AGC, 2 g of GC was first dissolved in 300 mL of distilled water (D.W.) and then mixed with 300 mL of methanol. Under continuous magnetic stirring, 16 mL of acetic anhydride was gradually added to the GC solution. HGCs were prepared by dissolving 3 g of GC in 375 mL of D.W. and subsequently diluting the solution with 375 mL of methanol. Hexanoic anhydrides (0.6204, 0.9926, 1.1167, and 1.2408 mL) were added drop-wise to a GC solution, respectively.

Following 48 h of stirring at room temperature, the polymers were precipitated with 1 L of acetone and isolated by centrifugation. They were then dialyzed in distilled water for 2 days to eliminate impurities, followed by freeze–drying lyophilization.

Characterization of AGC and HGCs

The chemical structures of AGC and HGCs were analyzed using ¹H NMR spectroscopy on an Avance III 600 spectrometer (Bruker, Germany) operating at 600 MHz. The polymer samples were prepared by dissolving them in D₂O at a concentration of 1% (w/v). The D₂O peak at 4.85 ppm served as the reference. The degrees of acetylation (DA) and hexanoylation (DH) were calculated by comparing the hydrogen peak of the glucopyranosyl ring (δ = 3.55 ppm) with that of the N-acylated substituent, using peak integration from the ¹H NMR spectra. The attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR) spectra of AGC and HGCs were obtained using a Nicolet iS5 spectrometer (Thermo Scientific, MA, USA) to analyze their structures. The analysis was conducted over 16 scans at a 4 cm⁻¹ resolution, covering a frequency range of 800–4000 cm⁻¹, to verify the chemical modifications.

Thermosensitive sol–gel transition

Thermo-reversible sol–gel transitions were assessed using a tube tilting technique, with the temperature raised at a rate of 0.1 °C per min. Each sample was prepared by mixing the polymer in D.W., PBS (pH 7.4), and RPMI-1640 at 4 °C to the required concentration. The gelation temperature was identified by tilting the tube and observing the absence of fluid movement after 30 s. The sol–gel transition phase diagram derived from this technique is accurate within ± 1 °C.

The gelation temperature was further evaluated using a rotating rheometer (TA Instruments, AR 1500ex, New Castle, DE, USA). Each polymer dissolved in RPMI-1640 was positioned between parallel plates with a 20 mm diameter and a 1 mm gap, applying a constant stress of 25 Pa. The sol–gel transition temperature was determined using a frequency sweep method at 1 Hz, with the temperature gradually increased at a rate of 0.16 °C/min, ranging from 10 to 45 °C.

Experimental animal

In this study, seven-week-old outbred Sprague–Dawley (SD) rats were obtained from Nara Biotech (Seoul, Republic of Korea). The animals were housed in a temperature-controlled room under specific pathogen-free (SPF) conditions, with a 12 h light/dark cycle. All animals were given unrestricted access to both water and food. All animal procedures were performed according to the Institutional Animal Care and Use Committee (IACUC) of Hanyang University (HY-IACUC- 2016-0043A).

Rat pancreatic islets isolation and culture

Male outbred Sprague Dawley (SD) rats, each weighing approximately 300 g, were utilized for the isolation of pancreatic islets. Collagenase P solution (1 mg/mL, Roche, Germany) was injected directly into the bile duct of each sacrificed rat. The swollen pancreas was then excised and incubated for 15 min in a 37 °C water bath for digestion. After digestion, the pancreases were washed twice with medium 199, filtered through a 450 μm sieve, and suspended in histopaque-1077. The suspension was centrifuged at 2,500 rpm for 17 min at 4 °C. Islets were collected from the gradient and further purified by gravity sedimentation. Purified islets were manually counted and cultured in RPMI-1640 medium containing 10% FBS and 1% penicillin–streptomycin. The islets were incubated overnight under standard culture conditions at 37 °C with 5% CO₂.

Preparation of single cells from pancreatic islets

Isolated islets (50–300 μm, 1000 IEQ) were washed 2 times with phosphate-buffered saline (PBS). Then, they were disrupted into single cells by mild enzymatic trypsin digestion (0.25%; trypsin–EDTA) at 37 °C for 10 min, and gently pipetted for 1 min. Trypsinization was terminated by adding RPMI-1640 medium supplemented with 10% FBS, and the resulting single-cell suspension was centrifuged at 1,800 rpm for 3 min. The collected single cells were washed with PBS 2 times for further use.

Preparation of gel and coating plates with AGC or HGC

Comparing spheroids' formation methods using AGC (4%, w/v), HGC20 (4%, w/v), HGC32 (3.5%, w/v), HGC36 (2%, w/v), and HGC40 (1.5%, w/v) were dissolved in RPMI-1640 with 10% FBS and 1% penicillin–streptomycin at room temperature. Next, each polymer solution (200 μL) was evenly distributed to the wells of a 24-well plate using a 1 mL syringe to coat the culture plates. The AGC- and HGCs-coated dishes were incubated at 37 °C for 1 h and then used to formation of spheroids for 2 days. In the case of gelation, each polymer solutions were mixed with islet single cells (1 × 10⁶ cells). The mixture was placed on a 24-well plate (SPL Life Sciences Co., Ltd., Seoul, Korea) and then incubated for 2 days at 37 °C with 5% CO₂. In the case of combination with coating and gelation, plate coating was first performed using AGC (4%, w/v). Then, the HGC36 (2%, w/v) solution encapsulated with islet single cells was cultured on top of it.

After 2 or 4 days of incubation, islet cell spheroids formed on various plates and gels were harvested by reducing the temperature to room temperature for 30 min and centrifugation (1,800 rpm, 3 min). During this process, non-aggregated cells were filtered out using a cell strainer (40 µm) (SPL Life Sciences Co., Ltd., Seoul, Korea) to separate the spheroids from the non-aggregated cells. The spheroid’s morphologies were examined under an optical microscope (Eclipse TE2000-S, Nikon, Japan) following incubation for 2 days and 4 days. To assess the size distribution of spheroids, images of each gel and culture plate were captured, and the spheroid count was analyzed using ImageJ software.

Cell viability analysis

Islet spheroids were incubated in a culture medium containing 10% CCK-8 solution at 37 °C with 5% CO₂ for 2 h. Absorbance of the medium was recorded at 450 nm with a SpectraMax M2 microplate reader. Subsequently, cell viability was adjusted based on the DNA content of the collected spheroids, which was quantified using the Pico Green DNA Assay Kit.

To assess cell viability in the islet spheroids, LIVE/DEAD staining was conducted. The spheroids were incubated in PBS containing 1 μM calcein-AM and 1 μM ethidium homodimer-1 (EthD-1) at 37 °C with 5% CO₂ for 20 min. Fluorescence microscopy was then used to capture the signals from live and dead cells.

Glucose-stimulated insulin secretion (GSIS) assay

To evaluate the functionality of islet spheroids, a glucose-stimulated insulin secretion (GSIS) assay was conducted. Islet spheroids and intact islets were initially incubated in Krebs–Ringer bicarbonate HEPES buffer supplemented 2.8 mM glucose for 30 min. Afterwards, the samples were then incubated in a low-glucose buffer (2.8 mM) or a high-glucose buffer (20.2 mM) for 2 h. Insulin concentrations in the medium were measured using a commercial rat insulin ELISA kit (Alpco Diagnostics, Salem, NH, USA). The stimulation index (SI) was determined by the ratio of insulin secretion under high glucose conditions to that under low glucose conditions, with both values normalized respective DNA values.

Statistical analysis

Data are presented as mean ± standard deviation (S.D.). Statistical comparisons between groups were performed using a two-tailed unpaired t-test to assess significant differences.

For the statistical analysis comparing multiple data groups, one-way analysis of variance (ANOVA) was employed, followed by Duncan’s multiple comparison tests. The Data are analyzed using the GraphPad Prism 8 program (La Jolla, CA, USA).

Results

Synthesis and characterization of AGC and HGCs

AGC and HGCs were produced via a straightforward one-step N-acylation process using carboxylic anhydrides, specifically acetic and hexanoic anhydrides (Fig. 2A) [18]. Characterization of the synthesized AGC and HGCs was conducted through ¹H-NMR and ATR-FTIR spectroscopy. The reference peak was established at 4.85 ppm, corresponding to the D₂O peak (Fig. 2B). In GC and AGC, peaks characteristic of the glucopyranosyl rings were detected in the range of 3.5–4.0 ppm, corresponding to the protons (H2 through H8). Furthermore, the peak at 2.6 ppm was assigned to the proton associated with primary amine residues (–NH₂). The acetyl group’s methyl protons were detected at 2.2 ppm, with this peak showing a significantly greater intensity in the N-acylated product compared to GC. The degree of acetylation (DA) for the synthesized AGC was determined by analyzing the integration of hydrogen peaks from the glucopyranosyl ring (δ = 3.55 ppm) to those from the N-acylated substituent, utilizing peak integration from ¹H NMR. The molar ratio of acetic anhydride to glucosamine residues in GC was adjusted to yield AGC with a degree of acetylation (DA) of around 91.5% (Table 1). The AGC’s chemical structure was also verified by ART-FTIR analysis (Fig. 2C). The absorption peaks detected at 2890 cm⁻¹ and 1596 cm⁻¹ corresponded to the − CH₂ groups and the amino group of GC, respectively. Absorption peaks at 1655 and 1555 cm⁻¹ were associated with the carbonyl stretching and bending vibrations of amide II of AGC, respectively. In the case of AGC, the amino vibration band at 1596 cm⁻¹ was absent, while the amide II band at 1555 cm⁻¹ showed increased intensity. The findings from the ¹H-NMR spectra and ATR-FTIR analysis aligned with previous studies, suggesting that AGC was successfully synthesized [18, 24].

Fig. 2.

Synthesis and characterization of AGCs and HGCs. A Chemical structure resulting from the N-acylation reaction of GC. B, C ¹H NMR spectra (B) and ATR-FTIR spectra (C) for GC and AGC. D, E ¹H NMR spectra (D) and ATR-FTIR spectra for GC and HGCs (E)

Table 1.

Characterization of AGC

Group	Feed molar ratioa	DAb (%)	Yield (%)
AGC	19	91.5 ± 0.8	80.4

^aFeed molar ratio of acetic anhydride to the GC’s the amino group

^bDegree of acetylation (DA) assessed through peak integration of ¹H NMR. Data are expressed as mean ± S.D. (n = 3)

The ¹H-NMR spectrum of all HGCs, based on the degree of hexanoylation (DH), displayed specific peaks at 0.89 ppm (–CH₂), attributed to the methyl protons, and at 1.199 ppm (–CH₂–CH₂–CH₃), 1.499 ppm (–CO–CH₂–CH₂), and 2.199 ppm (–CO–CH₂), corresponding to the methylene protons of the hexanoyl groups (Fig. 2D). The DH was determined to be approximately 20.6%, 32.3%, 36.5%, and 39.5%, respectively, by comparing the integration values of the proton peaks from the glucopyranosyl ring with those from the hexanoyl group (Table 2). The DH could be easily adjusted by altering the feed molar ratio of hexanoic anhydride, demonstrating a linear increase. The ATR-FTIR analysis revealed that the peak at 1655 cm⁻¹ is associated with the carbonyl stretching vibration, while the peak between 2850 and 2930 cm⁻¹ was attributed to the C–H stretching of methyl and methylene (Fig. 2E). Additionally, the peak at 1596 cm⁻¹, associated with the bending of amino groups in GC, nearly vanished, while new absorption peaks due to the bending vibrations of amide II bonds in HGCs appeared at 1555 cm⁻¹. These findings demonstrate that the N-hexanoylation of GC was successfully achieved [18, 24].

Table 2.

Characterization of HGCs

Group	Feed molar ratioa	DHb (%)	Yield (%)
HGC20	0.20	20.6 ± 0.8	83.3
HGC32	0.35	32.3 ± 0.8	82.5
HGC36	0.40	36.5 ± 0.9	80.3
HGC40	0.45	39.5 ± 0.4	83.8

aFeed molar ratio of hexanoyl anhydride to the GC’s glucosamine residue

bDegree of hexanoylation (DH) assessed through peak integration of ¹H NMR. Data are expressed as mean ± S.D. (n = 3)

Temperature-responsive sol–gel characteristics of AGC and HGCs

The temperature-responsive transition between the sol and gel phases of AGC and HGCs was examined through the tube tilting method and rheological analysis. Each polymer was diluted in distilled water (D.W.), PBS, and RPMI-1640 medium at concentrations of 1.5–5% (w/v) at 4 °C. The sol–gel phase transition was then evaluated by raising the temperature until reaching 50 °C. The temperature at which the sol–gel transition occurred was dependent on the solvent type and the DH. At the same polymer concentration, the phase transition occurred at a higher temperature in the following solvent types: D.W., PBS, and RPMI-1640 (Fig. 3A–E). However, we decided on RPMI-1640 as the solvent for AGC and HGC since RPMI provides a stable and physiologically relevant cell environment. Additionally, polymers with a higher DH transitioned to gel at lower concentrations. The phase transition occurred at a lower temperature as the polymer concentration increased in the same solvent. Notably, no sol–gel transition occurred below the specified concentration range, whereas higher concentrations led to unintended gelation at room temperature.

Fig. 3.

Properties of thermo-reversible AGC and HGCs. A–E Polymer concentration-dependent changes in gelation temperature of AGC (A), HGC20 (B), HGC32 (C), HGC36 (D), and HGC40 (E) dissolved in various solvents (1.5–5%, w/v). *: No gelation at the given concentration. F–K Temperature-dependent rheological behavior of GC (F), AGC (4%, w/v) (G), HGC20 (4%, w/v), (H) HGC32 (3.5%, w/v) (I), HGC36 (2%, w/v) (J), and HGC40 (1.5%, w/v) (K) dissolved in RPMI-1640. L Representative sol–gel transition images of GC, AGC (4%), HGC20 (4%), HGC32 (3.5%), HGC36 (2%), and HGC40 (1.5%) at 20 °C and 37 °C. The gelation behavior of each sample was observed by tilting the vials, demonstrating their sol or gel state

To achieve efficient thermogelling under physiological conditions and ensure the viability of islet single cells in RPMI-1640 medium, we determined the polymer concentrations as follows: AGC (4%, w/v), HGC20 (4%, w/v), HGC32 (3.5%, w/v), HGC36 (2%, w/v), and HGC40 (1.5%, w/v). The AGC and HGC’s thermosensitive gelation properties were evaluated using dynamic rheological methods over 10–45 °C (Fig. 3F–K). The gelation point (T_gel) was identified as the temperature at which the storage modulus (G′) and loss modulus (G′′) intersected during the evaluation of the sol–gel phase transition [18]. The observed negative thermo-responsive gelation is probably attributed to the formation of physical crosslinks among the hydrophobic acyl chains. Throughout the entire temperature range, the G′ values of the GC solution remained consistently lower than those of G’’ without exhibiting a crossover point, indicating that the GC solution did not undergo a thermo-sensitive sol–gel transition. In contrast, the G′ values for AGC and HGCs started off lower than G′′, but they significantly rose as the temperature increased. As a result, the sol–gel phase transition temperature was observed to be 36.3 ± 0.6 °C in AGC (4%, w/v) solution in RPMI, while the T_gel for HGC20 (4%, w/v), HGC32 (3.5%, w/v), HGC36 (2%, w/v), and HGC40 (1.5%, w/v) solutions were 30.3 ± 0.3 °C, 28.0 ± 1.0 °C, 36.2 ± 1.0 °C, and 31.7 ± 1.53 °C aligning with previous results obtained from the tube-tilting method (Fig. 3L).

Selection of optimal coating materials for islet cell spheroid formation

Simple methods for forming cell spheroids using non-adherent polymers include plate coating and gelation [20, 25, 26]. We aimed to identify the optimal method for spheroid formation by generating small and uniform islet cell spheroids using both approaches. First, we dispersed the pancreatic islet into single cells via trypsinization to make islet cell spheroids (Fig. S1A). Islet single cells were cultured on plates coated with AGC (4%), HGC20 (4%), HGC32 (3.5%), HGC36 (2%), and HGC40 (1.5%) for 2 days to determine the optimal coating materials for the formation of islet cell spheroids. Islet cell spheroids formed on coated plates were harvested by reducing the temperature to room temperature for 30 min. During this process, non-aggregated cells were filtered out using a cell strainer to separate the spheroids from the non-aggregated cells. As a result, we confirmed that each coating did not significantly affect the viability of islet single cells for 24 h (Fig. S1B). We observed that the islet single cells on the untreated dish were self-aggregated to a larger size than those on the coating plates (Fig. 4A) [27]. Among them, we observed that the AGC (4%)-coated plate facilitated the formation of the smallest and most uniform islet cell spheroids and significantly improved their yield (Fig. 4 B and C). It was also confirmed that a higher DH of HGC resulted in larger spheroid sizes and a lower yield of spheroids. The spheroids formed on the AGC (4%)-coated plate maintained their viability comparable to that of intact islets (Fig. 4D). Next, we conducted a glucose-stimulated insulin secretion (GSIS) test to evaluate the functionality of spheroids formed on each plate. The spheroids on the AGC (4%)-coated plate exhibited a similar insulin secretion pattern to the intact islet, and the stimulation index (SI), a marker of glucose responsiveness, was maintained (Fig. 4 E and F). Consequently, we determined that AGC (4%) was suitable for the coating materials for the formation of islet cell spheroids in terms of size and functionality.

Fig. 4.

Islet cell spheroids formed on the AGC and HGCs-coated plates. A The morphologies of islet cell spheroids formed on the coated plates after 2 days of incubation. Scale bar: 500 μm. B Size distribution of islet cell spheroids. Data are expressed as mean ± S.D. (n = 3). C Yield of islet cell spheroids. Yield was calculated as the number of spheroids formed on each plate relative to the initial islet single cell number and normalized to the untreated dish group. Data are expressed as mean ± S.D. (n = 3). Statistical significance was assessed using an unpaired, two-tailed t-test. D Viability of intact islets and spheroids after 2 days of incubation assessed by CCK-8 assay. Data are expressed as mean ± S.D. (n = 5). Statistical significance between groups was analyzed using one-way ANOVA followed by Duncan's multiple comparison test (p < 0.05). Groups labeled with different letters (a, b, c, and d) are significantly different, while groups sharing the same letter are not statistically different. E GSIS of intact islets and spheroids following 2 h incubation under low (2.8 mM) and high glucose (20.2 mM) conditions. Data are expressed as mean ± S.D. (n = 4). Statistical significance was assessed using an unpaired, two-tailed t-test. F SI of intact islets and spheroids assessed by GSIS assay. Data are expressed as mean ± S.D. (n = 4). Statistical significance between groups was analyzed using one-way ANOVA followed by Duncan's multiple comparison test (p < 0.05). Groups labeled with different letters (a, b, c, and d) are significantly different, while groups sharing the same letter are not statistically different

Selection of optimal gel materials for islet cell spheroid formation

Next, we cultured islet single cells on AGC and HGC gels for 2 days to identify the optimal gel materials for islet cell spheroid formation. We verified that none of the gels significantly affected the viability of islet single cells after 24 h (Fig. S2A) or the glucose responsiveness of intact islets after 24 h (Fig. S2B and S2C). Furthermore, we confirmed that islet single cells formed spheroids on each gel, and that higher HGC’s DH resulted in smaller spheroid sizes (Fig. 5 A and B). However, despite their small size, the spheroid yield on the HGC40 (1.5%) gel was similar to an untreated dish (Fig. 5C). Therefore, we selected HGC36 (2%) for gel materials, which produced small spheroids with high yield, as the next best option following HGC40 (1.5%). The spheroids formed on HGC36 (2%) gel maintained their viability, insulin secretion pattern, and SI comparable to those of intact islets (Fig. 5D–F).

Fig. 5.

Islet cell spheroids formed within the AGC and HGCs gel. A The morphologies of islet cell spheroids formed within the AGC and HGCs gel for 2 days of incubation. Scale bar: 500 μm. B Size distribution of islet cell spheroids. Data are expressed as mean ± S.D. (n = 3). C Yield of islet cell spheroids. Yield was calculated as the number of spheroids formed on each plate relative to the initial islet single cell number and normalized to the untreated dish group. Data are expressed mean ± S.D. (n = 3). Statistical significance was assessed using an unpaired, two-tailed t-test. D Viability of intact islets and spheroids after 2 days of incubation assessed by CCK-8 assay. Data are expressed as mean ± S.D. (n = 5). Statistical significance between groups was analyzed using one-way ANOVA followed by Duncan's multiple comparison test (p < 0.05). Groups labeled with different letters (a, b, and c) are significantly different, while groups sharing the same letter are not statistically different. E GSIS of intact islets and spheroids following 2 h incubation under low (2.8 mM) and high glucose (20.2 mM) conditions. Data are expressed as mean ± S.D. (n = 4). Statistical significance was assessed using an unpaired, two-tailed t-test. F SI of intact islets and spheroids assessed by GSIS assay. Data are expressed as mean ± S.D. (n = 4). Statistical significance between groups was analyzed using one-way ANOVA followed by Duncan's multiple comparison test (p < 0.05). Groups labeled with different letters (a, b, and c) are significantly different, while groups sharing the same letter are not statistically different

Effect of combining AGC coating and HGC gel on islet cell spheroid culture

Finally, we investigated the combined application of AGC (4%) coating and HGC36 (2%) gel to generate small and uniform islet cell spheroids through their synergistic effect. The plates were first coated using AGC (4%), followed by culturing HGC36 (2%)-encapsulated islet single cells for gelation on the coated plates at 37 °C for 4 days. As a result, the islet single cells in the combined treatment group successfully formed spheroids by day 2 (Fig. S3). By day 4, these spheroids exhibited the smallest and most uniform size compared to those formed under single treatments (AGC coating (4%): 200 μm, 27.5 ± 2.97%; HGC36 gel (2%): 200 μm, 30.7 ± 5.19%; AGC coating + HGC gel: 100 μm, 51.9 ± 6.72%) (Fig.6 A and B). Moreover, a high yield of islet spheroids was achieved under these conditions (Fig.6C). These spheroids exhibited significantly higher viability compared to intact islets and maintained sustained viability over 4 days, as indicated by a strong live cell signal and the absence of a dead cell signal (Fig. 6 D and E). In addition, they exhibited an insulin secretion pattern like that of intact islets (Fig.6F). Notably, the SI for these spheroids was superior to that of the intact islets. (Fig. 6G). These results demonstrate that the formation of smaller spheroids, compared to intact islets, significantly enhances both viability and glucose responsiveness [3].

Fig. 6.

Comparison of the viability and functionality of islet cell spheroids formed through the combined application of AGC coating and HGC36 gel. A The morphologies of spheroids on day 4. Scale bar: 200 μm B Size distribution of islet cell spheroids. Data are expressed as mean ± S.D. (n = 3). C Yield of islet cell spheroids. Yield was calculated as the number of spheroids formed on each plate relative to the initial islet single cell number and normalized to the untreated dish group. Data are expressed as mean ± S.D. (n = 3). Statistical significance was assessed using an unpaired, two-tailed t-test. D Viability of intact islets and spheroids after 4 days incubation assessed by CCK-8 assay. Data are expressed as mean ± S.D. (n = 5). Statistical significance between groups was analyzed using one-way ANOVA followed by Duncan's multiple comparison test (P < 0.05). Groups labeled with different letters (a, b, and c) are significantly different, while groups sharing the same letter are not statistically different. E Viability of spheroids formed on the combination plate coated with AGC (4%) and HGC36 (2%) gelation after 4 days incubation assessed by live/dead assay. Left: Bright field image. Right: Merged image. Green: live cells. Red: dead cells. Scale bar: 100 μm. F GSIS of intact islets and spheroids following 2 h incubation under low (2.8 mM) and high glucose (20.2 mM) conditions. Data are expressed as mean ± S.D. (n = 4). Statistical significance was assessed using an unpaired, two-tailed t-test. G SI of intact islets and spheroids assessed by GSIS assay. Data are expressed as mean ± S.D. (n = 4). Statistical significance between groups was analyzed using one-way ANOVA followed by Duncan's multiple comparison test (P < 0.05). Groups labeled with different letters (a, b, and c) are significantly different, while groups sharing the same letter are not statistically different

Effect of AGC coating and HGC36 gel on the intact islet culture

Intact islets were cultured on an AGC-coated (4%) plates, HGC36 gel (2%), and combination treatment (AGC coating + HGC36 gel) for 4 days to evaluate the effects of AGC coating and HGC36 gel on intact islet’s culture. As a result, the morphologies of intact islets on all groups did not show any difference compared to the untreated dish (Fig. 7A). We further verified that the AGC coating or HGC36 gel had no impact on the viability and functionality of the islets (Fig. 7B–D). As neither the AGC coating nor the HGC36 gel affected the viability or functionality of the islets, we suggested that the enhanced insulin secretion capacity of the islet spheroids formed through the combined application of AGC coating and HGC36 gel is attributed to their smaller size.

Fig. 7.

Intact islets cultured for 4 days on AGC-coated plates (4%), HGC36 gel (2%), and under the combined treatment conditions. A The morphologies of intact islets on day 2 and day 4. Scale bar: 200 μm. B Viability of intact islets after 4 days incubation assessed by CCK-8 assay. Data are expressed as mean ± S.D. (n = 5). Statistical analysis was performed using one-way ANOVA followed by Duncan's multiple comparison test (p < 0.05). All groups are labeled with the same letter, indicating that no statistically significant differences were observed among these groups. C GSIS of intact islets following 2 h incubation under low (2.8 mM) and high glucose (20.2 mM) conditions. Data are expressed as mean ± S.D. (n = 4). Statistical significance was assessed using an unpaired, two-tailed t-test. D SI of intact islets assessed by GSIS assay. Data are expressed as mean ± S.D. (n = 4). Statistical analysis was performed using one-way ANOVA followed by Duncan's multiple comparison test (p < 0.05). All groups are labeled with the same letter, indicating that no statistically significant differences were observed among these groups

Discussion

In this study, we developed and optimized a novel method for the formation of pancreatic islet spheroids using thermosensitive AGC and HGC as non-adhesive biomaterials. The varying gelation temperatures observed in different solvents (D.W., PBS, and RPMI-1640) can be attributed to the different interactions between the polymer chains and the solvents (Fig. 3). The higher ionic strength and osmolarity of PBS and RPMI-1640 can enhance interactions between polymer chains, promoting faster aggregation and resulting in gelation at lower temperatures [28, 29]. Additionally, as the polymer concentration increases, the proximity of polymer chains facilitates more rapid gelation, leading to lower phase transition temperatures [30]. The DH in HGC also played a significant role. Higher DH led to gelation at lower concentrations, likely due to increased hydrophobic interactions and a reduction in available reactive groups for cross-linking, which enhances the stability of the gel network [31, 32].

HGC involves introducing hexanoyl groups (C₆H₁₁CO-), which are hydrophobic chains, to the glycol chitosan molecule. This modification significantly increases the hydrophobic character of glycol chitosan by reducing the number of free amine groups available for hydrogen bonding with water and introducing hydrophobic interactions between the side chains [33]. Therefore, the hydrophobicity of HGC becomes more pronounced as its DH increases. In contrast, AGC involves introducing N-acetyl groups (–COCH₃), which replace the primary amine groups (-NH₂) in the glycol chitosan chain. Acetylation increases the formation of intermolecular hydrogen bonds and reduces the solubility of chitosan in water, but it does not introduce as strong hydrophobic interactions as longer fatty acid chains like hexanoyl groups [34, 35]. It was observed that coating with a polymer possessing strong hydrophobicity excessively enhanced cell–cell interactions, leading to an increase in the size of islet cell spheroids and a decrease in spheroid viability and glucose responsiveness (Fig. 4) [1, 2]. This indicates that AGC, with relatively weaker hydrophobicity than HGCs, is more suitable as a coating material for spheroid formation. Islet single cells are in direct contact with a non-adherent polymer surface on coated plates. The lack of a controlled environment led to random and non-uniform spheroid formation in untreated dishes [36]. In this context, the relatively weak hydrophobicity of AGC is adequate to maintain a balance by minimizing cell-substrate adhesion while still facilitating cell–cell interactions necessary for spheroid formation [20]. In contrast, islet single cells within the gel formed smaller islet cell spheroids as the hydrophobicity increased (Fig. 5). When cells are embedded within a gel, the environment becomes more complex due to the presence of a three-dimensional matrix that interacts with the cells in various ways [37, 38]. Higher hydrophobicity is required to ensure that cell–cell interactions dominate over cell–matrix adhesion, forming spheroids in a three-dimensional gel environment [39, 40]. While the high hydrophobicity of the HGC40 (1.5%) gel promotes compact spheroid formation, it can also reduce the diffusion of oxygen and nutrients into the spheroid. This is because the dense packing of cells creates diffusion barriers, leading to hypoxic conditions within the spheroid [41, 42]. Consequently, a decrease in cell viability and overall yield was observed. Therefore, this highlights the importance of balancing hydrophobicity in coating and gel materials to optimize the formation of islet cell spheroids. High-functional islet cell spheroids were successfully generated by combining AGC coating (4%) with HGC36 gel (2%) (Fig. 6). This improvement is likely due to the reduced susceptibility of smaller spheroids to hypoxia, which facilitates more efficient nutrients and oxygen diffusion [43, 44]. This approach produces smaller, more uniform, and functionally superior spheroids.

Our approach demonstrates several distinct advantages over conventional spheroid formation techniques, such as the hanging drop and mold-based methods. The hanging drop method often results in spheroids with a broad size distribution due to the manual nature of drop formation and the difficulty in controlling environmental conditions uniformly across drops [45]. Mold-based methods allow for precise control over spheroid size by using predefined geometries, ensuring uniform spheroids. However, the complex fabrication processes and high costs limit scalability, making these methods less practical for large-scale production [46, 47]. Our approach simplifies the process by leveraging the inherent properties of AGC and HGC to control spheroid size and uniformity without the need for complex equipment or multi-step procedures. This results in a more consistent and reproducible formation of spheroids with desired characteristics, making it more suitable for clinical applications. Another advantage of our method is its potential to be applied beyond islet cells. Although this study focused on islet single cells, the simplicity and tunability of HGC and AGC make it adaptable for forming spheroids from various cell types, including stem cells. This approach could benefit tissue engineering and regenerative medicine, where producing uniform, functional spheroids is critical for developing complex tissue structures.

Future research should explore the application of this method to other cell types to fully realize its potential. Additionally, in vivo studies are necessary to evaluate the therapeutic efficacy of the spheroids produced by this method in islet transplantation. The scalability of this method for large-scale production also remains an essential consideration for clinical translation.

In conclusion, the optimized method for spheroid formation using HGC and AGC presents a significant improvement over conventional techniques, offering a simpler, more efficient approach to producing smaller, more uniform, and functionally superior spheroids. This advancement holds promise not only for improving outcomes in islet transplantation but also for broader applications in cell-based therapies and tissue engineering.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

National Research Foundation of Korea (NRF), NRF-2022R1A4A1030421,DONG YUN LEE, National Research Foundation of Korea (NRF), NRF-2020R1A2C3005834, DONG YUN LEE, RS-2024-00449612, Kang Moo Huh, Korean Fund for Regenerative Medicine (KFRM), KFRM24A0105L1,DONG YUN LEE

Data availability

The datasets during and/or analyzed during the current study available from the corresponding author on reasonable request.

Declarations

Conflicts of interest

 The authors declare the following competing financial interest(s): Dong Yun Lee declares a financial interest in Elixir Pharmatech Inc., and Kang Moo Huh declares a financial interest in Lab-to-Lab Co. Ltd. These companies did not support the aforementioned research and currently has no rights to any technology or intellectual property developed as part of this research. The rest of the authors declare no conflict of interest.

Ethical statement

The animal studies were performed after receiving approval of the Institutional Animal Care and Use Committee (IACUC) in Hanyang University (IACUC approval No. HY-IACUC-2016-0043A).