Introduction

As per the International Diabetes Federation, 1 in 11 adults is living with diabetes (463 million people), which is expected to increase to 700 million by 2045, making diabetes a significant factor in premature mortality (1). Diabetes is not only a characteristic cause of abnormal metabolism in carbohydrates, fats, and proteins (2) leading to hyperglycemia (3) and hyperlipidemia (4) but also is closely associated with oxidative stress and inflammation (5) leading to microvascular complications such as retinopathy (6, 7), neuropathy (8,9,10), nephropathy (11), and macrovascular complications like coronary heart disease and stroke complications (12). Diabetes is commonly associated with chronic inflammation caused by monocytes’ enhanced production of inflammatory cytokines, which have an enhanced role in the pathogenesis of insulin (INS) resistance, impaired INS secretion, and impaired hepatic glucose metabolism (5, 13).

In type 2 diabetes (T2DM), the first phase of rapid biphasic INS secretion after glucose elevation is absent (14). The second phase diminishes with disease progression, resulting in the loss of pulsatile secretion leading to diabetes pathogenesis. Concomitantly, amplified oxidative stress and inflammation under diabetic conditions also lead to β-cell death, further exacerbating INS deficiency-related complications (15). T2DM progresses over time, and hyperglycemia is managed stepwise—starting with dietary restrictions and progressing to non-INS medication like sulfonylureas, biguanides, meglitinides, alpha-glucosidase inhibitors, and SGLT-2 inhibitors. As β-cell functioning declines, INS therapy may be necessary (16). Delivering INS systemically through the gut is difficult due to its instability and poor mucosal permeability; oral INS methods therefore need improvement. T2DM patients usually self-inject INS subcutaneously; however, knowledge of dose adjustments based on glucose monitoring is required to avoid infection, necrosis, and nerve damage. A broad range of delivery methods have been investigated for INS therapy (17), such as oral (18), pulmonary (19), nasal (20), suppositories (21), and transdermal approaches to alleviate the shortcomings of subcutaneous injections (22).

Hydrogels are three-dimensional polymer networks with high porosity that increase loading efficiency and release drugs over an extended period, improving efficacy and reducing toxicity and dosage requirements (23,24,25) thus emerging as a unique drug delivery method. Although various applications of hydrogels (23, 26) are gaining importance, stimuli-responsive injectable hydrogels have received additional attention due to minimal invasion, lack of organic solvents, less toxicity, and ability to deliver the drugs ranging from small molecules to large molecules like protein molecules.

The injectable chitosan gels, initiated by β-glycerophosphate salt at physiological temperature, invented by Ruel-Gariepy et al. and Chenite et al. (27, 28) gained emphasis due its improved bioadhesion, biocompatibility, antimicrobial properties, and texture matching tissue surface morphology making it an ideal choice for drug delivery, including loading large bioactive molecules like proteins (29,30,31). Hydrogels are well known for delivering hydrophilic drugs like INS (32, 33) due to the high-water content. For hydrophobic drugs, the hydrogel’s high-water content reduces loading capacity, mechanical strength, and stability. Copolymerization or conjugation with chitosan is therefore recommended for hydrophobic drug loading (34, 35).

We previously created a biosensitive hydrogel system that releases INS in response to elevated glucose levels. While the hydrogel indicated promising on-demand INS delivery, the inefficient INS loading (1 depot = 1 IU INS) complicated the tracking of INS release over multiple cycles (36). Very recently, we have demonstrated that the combination of long-acting injectable INS combined with oral nCUR is effective in delaying microvascular complications, compared to nCUR alone, while INS failed to offer protection at large (37,38,39). However, the combination regimens administered by two different routes does not seem to be compliant enough for translation. Literature evidence suggests that patients adhere to simple and less frequent dose regimens, and most importantly, combinations that can intervene multiple pathways of disease progression will be of significance (40, 41). However, combination drug products not only pose regulatory challenges but also make the formulation design difficult due to distinct physicochemical attributes of different drug classes (42). Herein, we report for the first time, glucose-responsive microgel (GRM), comprising hydrophilic drug (INS), and drug-laden nanoparticles (hydrophobic compound, curcumin) in the same formulation with a view to improve compliance. Overall, combination drug products such as this, combining traditional pharmaceutical drugs with food-grade natural anti-inflammatory agents, can be very effective in managing chronic diseases and associated complications.

Material and Methods

Materials

The materials used in this study are as follows: chitosan (85% deacetylated, Alfa Aesar, USA), catalase from a bovine liver or the peroxidase (POD) (Sigma-Aldrich), INS from bovine pancreas (Sigma-Aldrich), glucose oxidase (GOx, TCI, USA), gum arabic (GA) and curcumin (Acros Organics, USA), β-glycerol phosphate disodium salt hydrate (Chem-Impex International, Inc.), polyvinyl alcohol 88% hydrolyzed, Mw = 20,000–30,000 (Acros Organics), poly(lactide-co-glycolide) (PLGA)–Resomer 503H (Evonik, Germany) Mn = 30,000 g/mol (PLA:PGA; 50:50), and Quick Start™ Bradford protein assay (Bio-Rad). HPLC and LC–MS grade solvents were purchased from Fisher Scientific (USA). Distilled, deionized water was obtained from a Millipore Milli-Q purification system. FHs74 cell line (human small intestine epithelial cells), female of gestational age of 3 to 4 months, was obtained from ATCC, and passage 3 was used in the experiment.

Methods

Curcumin-Laden PLGA Nanoparticle (nCUR) Preparation

The curcumin-laden PLGA nanoparticle (nCUR) preparation, consisting of 0.12 mg/mg PLGA, has been optimized by our group (43,44,45,46,47). nCUR was prepared via an oil-in-water (O/W) emulsification and solvent evaporation method. In brief, PLGA (50 mg) and CUR (7.5 mg, 15% w/w) were dissolved in 2.5 mL ethyl acetate as an organic phase, while polyvinyl alcohol (PVA) (50 mg) was dissolved in 5 mL water as an aqueous phase, separately. After 30 min of dissolution, the oil phase was added dropwise to the water phase while stirring at 1500 rpm for 7 min. Subsequently, the emulsion was homogenized for 7 min at 15,000 rpm. The homogenized nanoemulsion was decanted into 20 mL of water, and the organic solvent was allowed to evaporate, followed by centrifugation at 15,000 g for 30 min at 4°C. The pellet was isolated and redispersed in a 6% aqueous sucrose solution and subjected to lyophilization. The entrapment efficiency of CUR in PLGA encapsulated nanoformulation was determined as 75% via HPLC.

Engineered Glucose Responsive Microgel (GRM) Comprising Insulin and nCUR

The gel preparation protocol was adapted from prior work in our lab (36). Briefly, a 2% chitosan (CS) solution was prepared by dissolving 100 mg of chitosan in 4 mL of an aqueous 1% HCl (v/v) solution and stirring vigorously at 1100 rpm until dissolved. To attain 7% (w/v) aqueous solution of β-glycerol phosphate disodium (GP), 0.5% of gum arabic (GA), glucose oxidase (GOx), and peroxidase (POD) aqueous solutions with a concentration of 0.4% (w/v) were prepared by dissolving 350 mg of GP in 250 µL water, 25 mg of GA in 250 µL water, and 20 mg of each GOx and POD enzyme in 200 µL of water. All percentages are calculated as per total volume of 5 mL batch. Upon continuous stirring of the CS solution placed in an ice bath, the solutions of GP, GA, GOx, and POD were added dropwise to CS solution, resulting in a pH range of 6.8–7.4 of solution. To prepare (nCUR) loaded GRM, lyophilized nCUR was dispersed in water (1 mg CUR/100 µL) and added dropwise to the abovementioned gel with an overall CUR concentration of 1 mg in the 5 mL solution. Aliquots of 200 µL of the solution were taken, and subsequently 5 IU of an aqueous INS solution (5 mg/mL, dissolved in 0.01 M HCl) was added. The mixture was thoroughly vortexed for 2 min and subsequently incubated at 37°C to form GRM microgel. Blank gel refers to only CS and GP, with no other additives.

Results

Preparation of GRM

The study aimed to find the optimum concentrations of thickening agent, added to chitosan and β-glycerol phosphate (GP) to create GRM and determining optimal conditions for high INS, enzyme loading. The optimized conditions for the GRM are 2% chitosan, 7% GP, 0.5% GA, and 0.4% GOx and peroxidase (POD), respectively. The total volume of the GRM sol was adjusted to 5 mL. The 200 µL of the sol was used every time to conduct an experiment in which 5 IU of bovine INS was added and thoroughly blended before gelation.

Optimizing Concentrations of Gum Arabic (GA), Enzyme (POD), and nCUR

Gum Arabic Variation

GA was added to the GRM to enhance viscosity and mechanical strength, which is crucial for accommodating higher enzyme payloads. Different concentrations of GA (0.3–2.5% w/v) were tested, and an optimum gel formation was achieved at 0.5% loading (Fig S1). Further increased GA concentration resulted in a thick sol. The concentrations of chitosan and GP are at 2% and 7%, respectively, and this does not include GOx and POD.

Optimization of POD Variation

We ensured that while attempting the loading efficiency, the gelation time is within < 10 min of injection so that an intact depot is formed enabling glucose-responsive nature. Various concentrations of POD were evaluated to obtain a uniform consistency of polymeric matrix, maintaining the activity and enzymatic stability. POD concentrations ranged from 0.0 to 0.4% (w/v), while gelling times remained consistent < 10 min. However, the concentrations of chitosan, GP, and GA are at 2%, 7%, and 0.5%, respectively, which remained constant.

In Vitro Release Studies

Release of INS in PBS Versus Glucose Solutions

In vitro release of INS through GRM in PBS showed an initial burst followed by sustained release until reaching a plateau (Fig. 1a). However, GRM placed in glucose solution presented a higher INS release indicating glucose responsiveness.

Fig. 1
figure 1

In vitro release of INS from GRM a in PBS and glucose (5 IU INS), b in 300 and 500 mg/dL of glucose (5 IU INS), and c with increasing concentration INS

Full size image

Release of INS with Varying Concentrations of Glucose

GRM showed on-demand INS release when placed in 300 mg/dL glucose concentration, along that of a diabetic patient. The INS release at 500 mg/dL glucose concentration gave broader peaks of insulin, ascribed to higher degrees of protonation in GRM leading to broader pore sizes, effectively displaying a larger INS release. However, a reduced number of cycles in pulsatile behavior were observed (Fig. 1b).

Release of INS with Varying Concentrations of INS Loading

We increased INS loading up to 5 IU/200 µL without compromising the integrity (36), gelling time, or stability of the gel. The INS release increased with higher loading up to 5 IU, but not with 7 IU, which resulted in longer gelling time and a soft gel. All other parameters were optimized based on the 5 IU loading (Fig. 1c).

Release of CUR

In the present study, we focused on attaining the synergistic effect of INS and CUR, where curcumin-laden nanoparticles (nCUR) were used in GRM. An initial burst for the first 2 h, followed by sustained release, was observed as measured by liquid chromatography-mass spectrometry (LC–MS) analysis (Fig. 2). The release of curcumin was continuous regardless of the medium used (water or glucose), indicating that the system’s gelling and release were independent of pH stimulus.

Fig. 2
figure 2

In vitro release behavior of GRM comprising nCUR

Full size image

Characterization of GRM

FTIR

FTIR analysis of the blank gel (CS + GP) and GRM in fresh and lyophilized form revealed several interactions (Fig. 3a). The broad peaks at 3400 cm−1 and 3600 cm−1 were assigned to N–H stretching and vibration of O–H in a hydrogen bond, while a peak at 1640 cm−1 was observed for C = O stretching of the amide bond, which is customary for INS, GOx, and POD in the GRM spectra. Additionally, the peak at 1650–1660 cm−1 was assigned to the α-helix structure of INS. The expected bands of polysaccharide and glycoprotein can be observed at 1377 cm−1 and a distinct band at 1029 cm−1 indicative of C-H from polysaccharide (36). The peak at 1400 and 1330 cm−1 for the bending vibrations of alkanes of GA (48), while the rise at 1750 cm−1 confirms the presence of C = O of PLGA in nCUR.

Fig. 3
figure 3

a FTIR spectra of GRM. b Loss factors of blank gel and GRM as a function of a temperature. c Complex viscosity as a function of temperature. d Syringeability of GRM in 37°C water. e Syringeability of GRM in 4°C water

Full size image

Rheology

To assess the elasticity of the GRM for pulsatile INS release, the blank gels and GRM were subjected to a heating and cooling cycle from 4 to 45°C. Loss factors as a function of forward and backward temperature cycles indicated no significant difference between blank gel and GRM, indicating very slow degradation of the polymeric network even in the presence of enzymes (Fig. 3b). The complex viscosity as a function of temperature was used to assess the mechanical/flow properties of the GRM. The addition of GA in the blank gel increased the complex viscosity, and smooth transitions of sol to gel were observed in GRM (Fig. 3c).

Syringeability and Injectability

Subcutaneous injections are typically recommended to be administered with 22–25 G needles and an injection volume of less than 1 mL, as these needles can penetrate up to 16–39 mm deep into the tissue (49). Studies show that high-viscosity injectables are well-tolerated, and injection volume and flow rate have minimal effects on pain perception in patients (50). We injected GRM using 23 G at 3 mL/h and observed phase transition when injected into 37°C water, forming gel depots at the bottom of the beaker (Fig. 3d). No gelation occurred when injected into 4°C water, which only resulted in a color change (Fig. 3e) (51). GRM met all five criteria for subcutaneous injection drug delivery as it was free flowing through 23 G and 25 G needles and had fast gelation for following initial burst release, biodegradability, efficient drug loading, and sustained release (52, 53).

Stability

INS Stability

Circular dichroism (CD) was used to evaluate the structural and conformational stability of INS released from the GRM and lyophilized GRM and compared to bovine INS solution. The CD spectra displayed characteristic peaks of INS at 208 and 222 nm, indicating that the secondary structure of INS was intact after gelation and lyophilization (Fig. 4).

Fig. 4
figure 4

CD spectra of INS, INS released from fresh and lyophilized GRM

Full size image

Lyophilization Effects

GRM is a thermoresponsive sol–gel that exhibits glucose-responsive behavior (Fig. 5a, b). However, it is a complex combination of enzymes, chitosan, GA, INS, and nCUR, making it susceptible to degradation. To preserve its integrity, lyophilization was performed, which protects sensitive biologics against environmental factors and increases shelf-life from days to months (54, 55). It was observed that GRM could be completely freeze-dried in a cake-like appearance (Fig. 5c). The nCUR was freeze-dried separately and later mixed with GRM at a concentration of 200 μg/mL. Upon resuspension of GRM and incubating at 37°C for regelation, a gel was formed without any significant change in gelation time compared to prior experiments with fresh GRM (Fig. 5d).

Fig. 5
figure 5

a GRM at room temperature. b GRM at 37°C. c Lyophilized GRM. d Gelation of lyophilized GRM

Full size image

Morphology

The SEM images of the blank gel and GRM display a smooth, three-dimensional microstructure with interconnected micropores, while the blank gel demonstrated average pore sizes of < 10 µm, whereas GRM displayed pore sizes of 1–20 µm. GRM indicated slightly larger pore sizes attributed by the additional components likely causing more phase separation (Fig. 6). The pore sizes play an important role in the drug release profiles of hydrogels, where a smaller pore size results in slower release, whereas a larger pore size leads to faster release (56). The tight networking of the GRM allows for slower release profiles compared to other chitosan hydrogels; however, release profiles will be environment (glucose)-dependent.

Fig. 6
figure 6

Representative scanning electron photomicrographs of blank and loaded GRMs

Full size image

Content Uniformity

GRM consists of blending various components, particularly two different active pharmaceutical ingredients (APIs). To ensure uniform drug content, we performed an HPLC content uniformity test for INS and CUR in both fresh and lyophilized GRM. Results indicated drug content variation within the acceptable range of 10% RSD, as shown in Table I.

Table I Content Uniformity for INS and CUR
Full size table

Cell Viability

The effects of GRM on cellular structure, proliferation, and viability of FHs74 cells were evaluated. Cellular integrity and structures were monitored in the presence of GRM (fresh versus lyophilized) upon staining actin filaments, which are important structural components of the cytoskeleton in eukaryotic cells. All cells in the presence of GRM indicated elongating structures, with no nuclear fragmentation observed (Fig. 7a). Additional background structures in the range of 1–5 µm were observed and ascribed to aggregated nCUR with supplemental experiments performed on GRM without nCUR. Additional BrDU labeling was performed after 24 h of co-culturing FHs74 cells (Fig. 7b) with GRM. In the presence of GRM, surviving cells were observed to proliferate, indirectly indicating minimal cellular death (Fig. 7c). Freshly prepared and lyophilized GRM did not show any differences in cell proliferation and viability.

Fig. 7
figure 7

a FHs74 cells and GRM (fresh vs. lyophilized), stained with actin and propidium iodide nuclear stain (scale bar = 10 µm), b with BrdU labeling (scale bar = 50 µm), and c BrdU positive cells per 650 µm2 determined from confocal microscopy

Full size image

Discussion

A variety of glucose responsive delivery systems were reported in the literature for on-demand insulin delivery (57,58,59,60), including chitosan-based systems (36). To the best of our knowledge, this is the first report describing the development of GRM-combined nCUR. The premise is based on our recent work where subcutaneous long-acting insulin combined with oral nCUR offered synergism in managing T1D complications (37,38,39). The goal in this present study was to combine both, insulin and curcumin in one formulation that can be self-administered and will be ideal to treat T2D and complications. The study aimed to determine optimal concentrations for additional components, including thickening agents, enzymes, and nCUR, added to a blank gel (consisting of chitosan and GP) to form GRM with improved INS loading capacity. GP is a fundamental component for creating thermoresponsive gels that utilize intermolecular interactions such as hydrogen bonding, electrostatic attraction between chitosan and GP via ammonium and phosphate groups, and hydrophobic interactions between chitosan molecules. To enhance INS loading capacity without compromising gelation time, various edible gums were evaluated to improve the mechanical properties of GRM. However, we observed that increasing INS loading did not accompany sol–gel transition and instead resulted in a liquid with restricted flow. GA, karaya gum, and xanthan gum were selected due to their binding and thickening properties. Karaya gum and xanthan gum did not provide the necessary viscosity upon increased INS loading. The high viscosity of xanthan gum may attribute to the failure, while karaya gum’s poor aqueous solubility consequently formed a viscous colloidal sol, even at low concentrations (61). In contrast, GA added the necessary mechanical strength for GRM stability. This could be due to its complex polymeric network of arabinose and galactose, which has a large hydrodynamic volume that can initiate complex intermolecular entanglements, even at low concentrations (62). Additionally, GA is an intrinsic part of the food industry as edible glue derived from gluten-free plant derivatives and thus well tolerated by most people, avoiding possible future toxicities during in vivo studies or translational process.

GOx acts as the sensor for the conversion of glucose to gluconic acid, lowering the pH, allowing the swelling in the pH-sensitive GRM, consequently generating an increase in pore size, and facilitating the release of the entrapped INS (36). However, the presence of POD acts as a scavenger of the H2O2 produced during the glucose-to-gluconic acid reaction. The absence of POD could result in toxicity or deactivation of GOx, leading to a disruption of the dynamic equilibrium in accordance with Le Chatelier’s principle. The cumulative release of INS is 15–20% higher in glucose than in PBS, providing evidence that generated gluconic acid protonates amine groups present on chitosan, in turn opening the pores in GRM, thus generating a higher INS release. In PBS, the GRM exhibits diffusion-driven release kinetics since there is no gluconic acid produced from glucose and thus no pH-responsiveness of GRM. The sustained release of INS was attributed to the natural disintegration of the gel over time when kept in the same PBS media. However, when media was intermittently replaced with water and glucose solution, an on–off pulsatile INS release profile was observed over 12 h that imitate the natural physiological conditions after and between meals. Upon increasing the glucose concentration from 300 to 500 mg/dL of the perfusion medium, INS displayed an irregular release pattern. INS release occurred in water (“fasting”) media, indicating that high glucose levels may be disintegrating the polymeric network, consequently releasing all possible available INS from the depot. However, in T2DM, a very high blood glucose level of 500 mg/dL is considered fatal. Thus, we can conclude that the GRM platform can withstand a sudden unexpected glucose spike in the system. After incorporating GA in GRM, we were able to increase the insulin loading capacity of GRM from 1 to 5 IU in comparison to our previous studies (36), as indicated in Fig. 1c, where INS release increased from 1 IU < 3 IU < 5 IU as expected. Better release profiles of 5 IU than 7 IU INS loading were achieved. With 7 IU/200 µL loading, the gel-forming ability and gelation time were significantly impacted. It is presumed if gelation time is extended after injection into the system, the sol form of GRM might get diluted with bodily fluid and disperse before forming a gel depot. For the prolonged INS release from the GRM depot, the removal of hydrogen peroxide is required. Insufficient POD restricts INS release in the following cycle (Fig S2), and excess can cause toxicity and accelerated chitosan polymeric network bulk erosion.

Our study demonstrates incorporating nCUR into chitosan hydrogels for drug delivery, resulting in multifaceted benefits. The benefits include on-demand INS supply and sustained release of CUR, which can protect against chronic side effects of T2DM such as inflammation. From a translational aspect, a subcutaneous injection combining two compounds of distinct physicochemical attributes in one will reduce the daily number of insulin administrations with additional benefit of CUR.

Another important aspect of the characterization of hydrogels is their rheological properties. Complex viscosity is a measurement of a material’s total resistance toward the flow. Thus, constant and smooth increments of complex viscosity are always an indicator of the thermal stability of gel (63). A similar behavior was observed for GRMs, indicative of uniform crosslinking alignment in polymeric structure. We have observed that in the absence of GA to GRM, a decrease in the viscosity is observed and the sol did not undergo gelation. As indicated in Fig. 3c, GA provided the required viscosity, which was sufficient for gelation.

As GRM has the potential for translation, it is crucial to conduct stability studies. CD spectra confirmed that INS’s secondary structure and activity were not compromised during gelation or temperature variation during freeze-drying. After lyophilization, INS maintained its secondary structure without any disorder in the three-dimensional structure that could cause product disbarment during clinical trials (64). After lyophilization, resuspending the gel resulted in clumps due to nCUR disrupting the process. This may be due to the lack of additional cryoprotectants during lyophilization, resulting in nCUR undergoing irreversible aggregation and losing its moisture-absorbing capacity when rehydrated. As a result, the nCUR shrunk and was unable to fill the voids in the chitosan and glycerol phosphate network, causing the gel to form clumps instead of gelation occurring (65). To overcome nCUR disturbing the resuspension of lyophilized GRM, it was removed and freeze-dried separately and subsequently re-dispersed individually and mixed in the desired concentration of 40 µg/200 µL of GRM gel depot resulting in a uniform sol.

Conclusion

The GRMs underwent detailed characterization to allow for storage without affecting the gelling properties. In vitro studies showed that GRMs had a higher INS loading capacity than GA-free gels. The release rate of INS was affected by its loading and glucose concentration, while nCUR was released at a constant rate regardless of glucose concentration. These results suggest that GRMs can release INS on demand and sustain CUR release. The in vitro study conducted on the FHS74 cell line demonstrated that GRMs in its formulated or lyophilized state did not cause undesired effects, thus making it a safe option for further testing of this combination in vivo which may be useful in treating T2DM and its complications.