Introduction

In a rapidly aging society with declining birth rates, the urgent need for effective bone regeneration methods is increasingly apparent. According to a World Health Organization report, the global population aged over 60 years is expected to nearly double from 12% to 22% between 2015 and 2050. This demographic shift contributes to an increase in patients experiencing bone loss due to conditions such as osteoporosis [1], osteoarthritis [2], and rheumatoid arthritis [3]. Common clinical methods for repairing bone defects include bone grafting [4,5,6,7], the Masquelet technique [8], and the Ilizarov technique [9]; however, these methods involve high surgical invasiveness and a significant patient burden. Alternatively, bone tissue engineering (BTE) utilizes scaffold material to load cultivated seed cells in vitro and place them into the defect site to achieve bone healing [10]. Therefore, for successful bone regeneration, an optimal interaction between the scaffold and osteogenic cells is crucial and warrants further exploration.

Over the past few years, hydrogels with three-dimensional crosslinked polymer chains have garnered attention in regenerative therapy. These hydrogels, which mimic the properties of biological tissues and possess high water content, can be modified with various functional molecules. Typically, when hydrophilic polymers such as polyvinyl alcohol (PVA) [11, 12] or polyethylene glycol (PEG) [13] are used, these hydrogels are hydrophilic and do not have mesoporous structures. However, ideal scaffolding materials should exhibit mesoporous structures with balanced hydrophilic and hydrophobic properties, mimicking the extracellular matrix to provide mechanical stability and promote cell functions.

Recently, we discovered gel–gel phase separation (GGPS) in PEG hydrogels constructed solely from a 4-arm PEG polymer. This process not only results in the formation of a mesoporous structure but also modifies the hydrophilicity of PEG, increasing its hydrophobicity [14, 15]. Consequently, hydrophobic microparticles adhered to the phase-separated PEG hydrogel surface, demonstrating altered surface properties due to hydrophobic interactions. In vivo, phase-separated PEG hydrogels promoted adipose-like tissue formation when subcutaneously implanted in rats; phase-separated PEG hydrogels also demonstrated slow release of bone morphogenetic protein-2 (BMP-2), enhancing new bone formation in mouse cranial defect models [14, 16]. These findings suggest that the hydrophobization induced by GGPS can transform the original properties of polymer hydrogels and impart new characteristics.

However, further optimization is necessary to fully exploit the potential of this phenomenon. One crucial aspect that has not been extensively investigated in GGPS is the branching factor of the hydrogel network. We hypothesize that increasing the branching factor from 4 to 8 and incorporating RGD (Arg-Gly-Asp) peptides, which are cell adhesion peptides, will increase the cell affinity of PEG hydrogels. Although our previous studies demonstrated that phase-separated PEG hydrogels have biological affinity in vivo, they are still biologically inert in vitro [14, 17]. Anchoring RGD to the hydrogel network can reduce the immune response and allow directional binding modification, minimizing the impact on the physical properties of the base materials [18, 19]. This approach could clarify the impact of GGPS on cells and advance its application in regenerative therapy.

Compared with 4-arm PEG, the use of PEG with a greater number of branches, such as 8-arm PEG, can enhance the cross-linking efficiency under identical conditions. The 8-arm PEG facilitates a greater extent of modification, accelerates the reaction rate, reduces swelling, and yields a more suitable network structure, all while requiring only an equivalent molar concentration of the reactive functional groups [20, 21]. Therefore, in this study, we conducted a comparative analysis of 4-arm and 8-arm PEG hydrogels and investigated the effect of GGPS on osteogenic differentiation in vitro by introducing RGD peptides. By focusing on the potential of increased branching factor and RGD incorporation to increase cell affinity, this study provides insights into the chemical and biological properties of PEG hydrogel scaffolds for bone regeneration.

Methods

Materials

The polyethylene glycol maleimide (PEG-MA, Mw = 10 kDa) and polyethylene glycol sulfhydryl (PEG-SH, Mw = 10 kDa) were purchased from SINOPEG (Xiamen, China). The types of multiarm PEG-MA and PEG-SH included 4-arm (4-arm PEG-MA Lot No. P2211070190901-B, 4-arm PEG-SH Lot No. P2205070130902-B) and 8-arm (8-arm PEG-MA Lot No. YP2106090190901-B, 8-arm PEG-SH Lot No. YP2106090130901-B). The citric acid and disodium hydrogen phosphate used as buffer components were obtained from FUJIFILM (Tokyo, Japan). Pure water was obtained via Milli-Q Synthesis (Vent Filter MPK01) and purchased from MERCK (Tokyo, Japan). All the materials were used without further purification.

Sol–gel phase diagram

Citrate phosphate buffer (CPB) with a 50 mM salt concentration and various pH values (5.8, 5.0, 4.6, 3.8, and 2.6) was made by adjusting the mixing proportions of citric acid and disodium hydrogen phosphate. PEG-MA was dissolved in CPB at various PEG concentrations depending on the pH (c = 10–100 g/L) (c < 10 g/L at pH 5.8, 10 g/L < c < 30 g/L at pH 5.0, 30 g/L < c < 80 g/L at pH 4.6, 80 g/L < c < 100 g/L at pH 3.8, and c > 100 g/L at pH 2.6). PEG-SH precursor solutions were prepared via the same procedure. The PEG-MA and PEG-SH precursor solutions with the same c were mixed at various mixing ratios, r = [PEG-SH]/([PEG-SH] + [PEG-MA]) (r = 0.01–0.90), at which the hydrogel form was determined by adjusting the molar ratio of the prepolymers.

In accordance with this process, 8-arm-PEG-MA and 8-arm-PEG-SH precursor solutions were mixed at different r ratios to make 8-arm PEG hydrogels with a total volume of 100 µL. The range of r was a minimum of 0.01, increasing stepwise by 0.05 to a maximum of 0.90. The mixed solution was added to a 1.5 mL Eppendorf tube. All hydrogel samples were maintained at room temperature for 24 h for complete gelation. The sol–gel transition was tested and recorded via macroscopic observation of inverted tubes. The 4-arm PEG hydrogel was evaluated via the same procedure.

Preparation of RGD-modified PEG hydrogels

PEG-SH and PEG-MA were each dissolved in CPB at pH 5.0 and then sterilized via a 0.22 µm filter. The RGD (GRGDSPC) solution was added to the PEG-MA solution at concentrations of 0, 0.4, and 0.8 mM. and incubated at 25 °C for 5 min. Then, RGD-PEG-MA and PEG-SH were mixed at c = 20 g/L under stoichiometrically balanced conditions (r = 0.5) to form the RGD-PEG hydrogel.

Rheological analysis of hydrogels

The PEG hydrogels were poured into the gap within the rheometer double cylinder. The time-dependent changes in the storage modulus (G’) and the loss modulus (G”) were measured with a double-cylinder geometry rheometer (MCR302, Anton Paar, Austria) with an oscillating frequency of ω = 0.1–10 rad/s at 25 °C.

Turbidity of hydrogels

The PEG hydrogel precursors were mixed at different c or r values and poured into plastic cells with an optical length of 5 mm. The transmittance at λ = 400 nm was measured via a UV‒vis spectrophotometer at 25 °C. The turbidity (D) was estimated by normalizing the absorbance to the polymer concentration (g/L).

Degree of swelling

The gelation solution was poured into a cylindrical capillary glass (2 mm in diameter) (Hirschmann Ringcaps®). The hydrogel was subsequently carefully extracted after breaking the capillary glass, and the diameter at this point was defined as d0. All the samples were immersed in water or PBS and incubated at 37 °C for 1 wk, after which the diameter was called d. The diameters were evaluated daily via an optical microscope. The degree of swelling was then calculated as Q = (d/d0)3.

Particle adhesion on the hydrogel surface

A polytetrafluoroethylene (PTFE) mold (15 mm diameter, 7 mm height) was used to prepare the hydrogels. The 8-arm PEG hydrogels and 4-arm PEG hydrogels were incubated for 24 h at 25 °C. The hydrogels were then removed from the mold and immersed in a 0.1% w/v solution of Fluoresbrite® YG carboxylate microspheres (10 μm diameter) in water for 24 h at 25 °C. Images were obtained via an LSM 800 confocal laser microscope (Zeiss, Japan), and the number of particles was quantified via ImageJ software.

Contact angle measurement

A PTFE mold (7 mm diameter, 3 mm height) was used to evaluate the contact angle. First, the hydrogels were immersed in water for 24 h at 25 °C. The hydrogel samples were then placed on a glass slide and a contact angle meter (DMs-401, Kyowa Interface Science Co., Ltd.). One microliter of water was dropped gently onto the gel surface by slowly moving the gel sample shelf. Contact angle measurements and images were obtained via ImageJ and FAMAS contact angle measurement software (Kyowa Interface Science Co., Ltd.).

Scanning electron microscopy

The hydrogel surface morphological images were obtained via a TM3030 scanning electron microscope (Hitachi, Tokyo, Japan) under EDX conditions. The hydrogels (15 mm in diameter and 3 mm in height) were immersed in water for 2 h before imaging. The microscope is capable of imaging hydrogels in a wet state.

Cell culture

The RGD-PEG hydrogels were immersed in PBS and incubated at 25 °C for 24 h before the cell culture experiments were conducted. MC3T3-E1 cells were obtained from the Riken Cell Bank (Tsukuba, Japan) and maintained in alpha-modified Eagle minimum essential medium (α-MEM) (11900-073. Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS) (Lot. M302683268-1; Equitech-Bio), and 1% penicillin/streptomycin (Cat. No. P4333; Sigma‒Aldrich) and were cultured in 100 mm petri dishes at 37 °C with 5% CO2. The culture medium was replaced every two or three days. The cells were passaged via 0.25% trypsin/EDTA until they reached 80–90% confluence.

Cell proliferation assay

A cell counting kit-8 (CCK-8) (Dojindo, Japan) was used to assess cell proliferation. MC3T3-E1 cells (1 × 104 cells/well) were seeded on hydrogels in 96-well plates. After the supernatants were removed, CCK-8 solution was added to each well for 1 h at 37 °C, and the absorbance at 450 nm was measured via a Multiskan Spectrum Microplate Spectrophotometer.

Cell differentiation assay

MC3T3-E1 cells (2 × 105 cells per well) were cultured in a 12-well plate with or without hydrogel treatment until reaching 80–90% confluence, after which they were changed and maintained in osteogenic medium consisting of high-glucose Dulbecco’s modified Eagle’s medium (D-MEM) (043–30085, FUJIFILM Wako) supplemented with 10% FBS, 1% penicillin/streptomycin, 50 µg/mL ascorbic acid (012–04802, FUJIFILM Wako), 10 mM β-glycerophosphate (G9422–100G, Sigma–Aldrich), 0.01 µM dexamethasone (041–18861, FUJIFILM Wako), and 100 ng/mL recombinant human bone morphogenetic protein-2 (rhBMP-2) (7510050, INFUSETM Bone Graft, Medtronic). The osteogenic medium was replaced every two or three days.

RT‒PCR

Gene expression levels were detected via quantitative real-time RT‒PCR as described previously [16]. Total RNA was extracted from cells via ISOGEN (311-02501; Nippon Gene) and an RNeasy Mini Kit (74106; Qiagen). The quality and quantity of RNA from each sample were evaluated via a NanoDrop® ND-1000 spectrophotometer (Thermo Fisher Scientific). Next, 1 µg of total mRNA was reverse transcribed into complementary DNA (cDNA) via a ReverTra AceTM qPCR RT Master Mix with a gDNA Remover Kit (FSQ-301, Toyobo Co., Ltd.). Two microliters of cDNA was then amplified with specific primers for each reaction with a FastStart Universal SYBR Green Master Kit (04913 850001, Roche). RT‒qPCR was performed via a 7500 Fast Real-Time PCR System (Applied Biosystems). Actb was used as a mouse endogenous reference for the ∆∆CT method to evaluate relative gene expression. The following primers were used:

Mouse actin beta, Actb:

forward 5’-AGATGTGGATCAGCAAGCAG-3’,

reverse 5’- GCGCAAGTTAGGTTTTGTCA-3’.

Mouse alkaline phosphatase, Alpl:

forward 5’-GCTGATCATTCCCACGTTTT-3’,

reverse 5’-CTGGGCCTGGTAGTTGTTGT-3’.

Mouse integrin-binding sialoprotein, Ibsp:

forward 5’-CAGAGGAGGCAAGCGTCACT-3’,

reverse 5’-CTGTCTGGGTGCCAACACTG-3’.

Results and discussion

Sol–gel diagram

According to our previous findings, GGPS occurs at polymer concentrations just above the lowest concentration that forms a hydrogel [14]. To determine the optimal concentration for forming hydrogels with GGPS, we first investigated the sol–gel diagrams for both the 4-arm and 8-arm PEG hydrogels. The hydrogels were formed from mutually reactive PEG prepolymers comprising terminally reactive sulfhydryl (SH) and maleimide (MA) groups, which form thioether bonds [22]. This reaction proceeds nearly completely, exceeding 90%, even at low polymer concentrations [23]. Therefore, in the process of AB-type polymerization, it is necessary to consider the initial polymer concentration (c) of the PEG prepolymers. Additionally, the mixing ratio (r) is another crucial parameter that can be used to adjust the molar ratio of PEG prepolymers at each c. Here, r was expressed as r = [PEG-SH]/([PEG-SH] + [PEG-MA]). In our experiments, we modulated the r values of two prepolymers at various concentrations and determined the critical mixing ratio (rc) required for the formation of different branching PEG hydrogels via the tilting method (Fig. 1a). The obtained rc values confirmed that gelation occurred even under nonequimolar mixing conditions at relatively high concentrations. As shown in Fig. 1b, the boundary between the sol and gel states forms a U-shaped curve rotated 90 degrees counterclockwise, with the gel phase inside and the sol phase outside. As the polymer concentration decreases, hydrogel formation occurs only near equimolar conditions, reflecting the difficulty in forming hydrogels at lower concentrations. Interestingly, when comparing the 4-arm and 8-arm PEG systems, the two U-shaped curves cross at 40 g/L. At higher concentrations (40–100 g/L), the gelation ability of 8-arm PEG is greater than that of 4-arm PEG, which is reasonable because higher branching numbers theoretically enhance the gelation ability. Indeed, rc appears to asymptotically approach the Bethe lattice prediction (rc = 0.1667 and 0.071 for 4- and 8-arms, respectively) [24]. However, the behavior at lower concentrations is counterintuitive: the gelation ability decreases with increasing branching number. Moreover, the lowest concentrations forming hydrogels are c = 8 g/L in the 4-arm PEG system and c = 11 g/L in the 8-arm PEG system. This counterintuitive behavior is most likely caused by the difference in overlap concentration (c*), which is the polymer concentration required to saturate the system. As the c* of 8-arm PEG is smaller than that of 4-arm PEG [25, 26], the 8-arm PEG system is more dilute than the 4-arm PEG system at a given concentration. When we consider the difference in c*, the lowest concentrations forming hydrogels are almost identical for both systems (c/c* = 0.17).

Fig. 1
figure 1

Phase diagram and GGPS evaluation. a Schematic processes for preparing 4-arm PEG hydrogels (left) and 8-arm PEG hydrogels (right). The terminal Groups A and B were maleimide (MA) and thiol (SH), respectively. The precursors were mixed under different initial polymer concentrations (c) or different mixing ratios (r), r = [PEG-SH]/([PEG-SH] + [PEG-MA]). b Sol–gel phase diagram of the 4-arm PEG hydrogel and 8-arm PEG hydrogel depicting the relationship between c and r. Circles represent the critical gel state, and arrows represent the critical sol state. All the experiments were performed in triplicate. c In the glass bottles, at c = 20 g/L and r = 0.5, the appearance of the gels became turbid due to phase separation. d GGPS evaluation by turbidity D measurements of 4-arm PEG and 8-arm PEG hydrogels at nominalized concentrations, where c is the initial concentration and c* is the overlap concentration. e GGPS evaluation by turbidity D measurements of different mixing ratio r of 20 g/L 8-arm PEG hydrogel. All the experiments were performed in triplicate. The standard deviations are smaller than the sizes of the symbols

Comparison of 4-arm and 8-arm hydrogels

We investigated the turbidity of the hydrogels after the gelation reaction was complete via a UV‒vis spectrophotometer, as turbidity is a precursor of GGPS [14]. In both hydrogels, higher turbidity was observed at dilute concentrations (c/c* < 1.0), whereas turbidity was not noticeable at semidilute concentrations (c/c* > 1.0) (Fig. 1d). This behavior is consistent with our previous results [14]. Notably, the turbidity of the 8-arm PEG hydrogels was approximately six times greater than that of the 4-arm PEG hydrogels at c/c* = 0.43, likely due to the increased network strand density in the 8-arm PEG hydrogel. Turbidity is a principal element of GGPS [27]. This result suggests that GGPS is more pronounced in 8-arm PEG than in 4-arm PEG. This expectation is further supported by the decreased turbidity observed under nonequimolar mixing of 8-arm PEGs (Fig. 1e). Under nonequimolar mixing conditions, the fraction of connected arms and the resulting network strand density decrease. These results strongly indicate that the increased strand density at lower polymer concentrations induces strong turbidity, which is a precursory observation of GGPS. Based on the turbidity results, we focused on the hydrogels formed at c = 20 g/L, as the turbidity is significant and both systems can stably form hydrogels. At this concentration, the 4-arm PEG hydrogel appeared almost transparent, whereas the 8-arm PEG hydrogel exhibited white turbidity (Fig. 1c).

We then compared the storage moduli (G’) of these two hydrogels measured by a rheometer with a double-cylinder geometry (Fig. 2a). In both hydrogels, the loss modulus (G”) was much lower than the G’ value; thus, the two hydrogels are almost elastic. The elastic moduli (G’ ≈ 500 Pa for the 8-arm hydrogel and G’ ≈ 1000 Pa for the 4-arm hydrogel) of both PEG hydrogels were suitable for use as artificial extracellular matrices [26]. These values are close to those of type I collagen found in bone ECMs. Type I collagen is a standard material used in bone regeneration because of its significant role in bone mineralization. In the bone ECM, type I collagen is one of the most abundant proteins, with osteoblasts secreting type I collagen, which accounts for 85–90% of the total protein in bone. This collagen plays a crucial role in bone mineralization [27]. The similarity in elastic moduli suggests that the PEG hydrogels can provide a supportive environment for cell adhesion, proliferation, and differentiation, making them suitable for applications in bone tissue engineering. However, this result contradicts the expectation that 8-arm PEG would exhibit a higher elastic modulus than 4-arm PEG because of its increased branching factor, which creates a denser network structure [20, 28]. This discrepancy may be explained by the increased turbidity of the 8-arm PEG hydrogel. As the branching number is greater in 8-arm PEG, it should form a denser network than 4-arm PEG locally. However, owing to the strong heterogeneity, as evidenced by the turbidity, the macroscopic modulus may not simply reflect the local structure due to heterogeneity. Conversely, the enhanced turbidity of the 8-arm PEG hydrogel is probably due to the denser network locally formed by the increased branching factor.

Fig. 2
figure 2

Physical property characterization of the 4-arm and 8-arm PEG hydrogels. a Elasticity of 4-arm and 8-arm hydrogels after the respective PEG precursors. The storage modulus (G’) and loss modulus (G”) of the 4-arm and 8-arm PEG hydrogels with 20 g/L, r = 0.5, were measured via the double-cylinder geometry method at 25 °C. b Swelling of the 4-arm and 8-arm PEG hydrogels after immersion in PBS at 37 °C. All the experiments were performed in triplicate. The standard deviations are smaller than the sizes of the symbols. c SEM surface morphology images of the surfaces of the 4-arm and 8-arm PEG hydrogels. Scale bars, 100 µm. d Adsorption of fluorescent hydrophobic microparticles to the hydrogel surface. The top row of images shows the whole view of the gels with scale bars at 5 mm, and the bottom row of images provides a magnified view with scale bars at 100 µm. The particle numbers were quantified via ImageJ. The data are presented as the mean ± SD, and n = 3 in each group. Statistical analysis was performed via Student’s t test. ****P < 0.0001. e Contact angles of the 4-arm and 8-arm PEG hydrogels determined via the sessile drop method. Yellow arrows denote the angle measured. Angle data were analyzed via ImageJ. Data are presented as the mean ± SD, and n = 3 in each group. Statistical analysis was performed via Student’s t test. **P < 0.01

Hydrophobic behavior of hydrogels with GGPS

Our previous research indicated that GGPS does not occur during the gelation process but does occur after the completion of the gelation process, accompanied by deswelling. Generally, the equilibrium condition of an electrically neutral hydrogel in water is determined by the balance of the repulsive osmotic pressure and the attractive elastic pressure [29]. The two hydrogels were removed from the molds, soaked in PBS, and incubated at 37 °C for 1 week. The degree of swelling (Q), calculated by comparing the volumes before and after immersion, decreased in both the 4-arm and 8-arm PEG hydrogels (Fig. 2b) and almost reached equilibrium within one day. Notably, conventional hydrogels with relatively high polymer concentrations swell [30], but the 4-arm and 8-arm hydrogels prepared here deswell. The observed deswelling of the PEG hydrogels at low concentrations was likely due to the elastic pressure exceeding the osmotic pressure [31]. The enhanced deswelling of the 4-arm hydrogel is consistent with its increased G’ compared with that of the 8-arm hydrogel. After reaching shrinking equilibrium, we carefully prepared cryosamples of these hydrogels and observed them using scanning electron microscopy (Fig. 2c). While no characteristic structure was observed on the surface of the 4-arm PEG hydrogel, a mesoscale filamentous structure was observed only on the surface of the 8-arm PEG hydrogel (Fig. 2c), indicating that the 8-arm hydrogel caused stronger GGPS than did the 4-arm hydrogel. These observations help us understand the mechanism of GGPS. Previously, increased turbidity (local network density), increased elastic modulus, and enhanced deswelling were considered indicators of GGPS [32]. These results strongly suggest that local network density rather than the macroscopic elastic modulus or deswelling governs GGPS. Therefore, even if 8-arm PEG 20k and 4-arm PEG 10k were used to equalize the arm length, the hydrogel with a higher local mesh density would still exhibit greater turbidity, resulting in increased hydrophobicity. This finding indicates that 8-arm PEG, with its increased chain density, is more prone to phase separation, leading to greater turbidity and enhanced hydrophobicity. Notably, unlike typical phase separation behaviors induced by adding salts, other polymers, or high polymer concentrations [33], we prepared phase-separated hydrogels simply by forming hydrogels near the critical gelation concentration without these elements. This unusual shrinking behavior is also advantageous, as gel swelling in vivo can cause pressure on the surrounding tissues, leading to undesired tissue damage.

Unexpected hydrophobicity is another intriguing feature of GGPS hydrogels [14, 16]. To investigate the adsorption of hydrophobic particles on the hydrogel surface, we immersed the deswollen hydrogels in a suspension of hydrophobic particles. Remarkably, the 8-arm PEG hydrogel exhibited greater particle adsorption than did the 4-arm PEG hydrogel, with approximately 4 times more adsorbed particles (Fig. 2d). This observation suggested that the hydrogel surface became hydrophobic, physically adsorbing the particles through hydrophobic interactions. This phenomenon is reminiscent of the dehydration-induced hydrophobization observed in polyN-isopropylacrylamide (PNIPAM) above its lower critical solution temperature [34, 35]. The increased hydrophobicity of the 8-arm PEG hydrogel indicates that increasing the branching number can increase the hydrophobicity caused by GGPS. Interestingly, the particles, with a diameter (≈10 µm) much smaller than the characteristic size of phase separation (≈100 µm), did not permeate into the hydrogel. This observation suggests the absence of “macro pores” in the hydrogels with GGPS, where two distinct gel phases, one dilute and one dense, coexist. To further corroborate the hydrophobicity, we measured the water contact angle on the hydrogel surface. Compared with the 4-arm PEG hydrogel, the 8-arm PEG hydrogel presented a greater contact angle (Fig. 2e). Our previous work suggested that surface wettability is independent of the PEG network structure and only weakly depends on surface charge [36]. However, since the 4-arm and 8-arm PEG hydrogels investigated here do not possess noticeable surface charges, the effect of surface charge can be ruled out. Consequently, the enhanced hydrophobicity can be attributed solely to the enhanced hydrophobization caused by GGPS.

Cell adhesion behavior in RGD-PEG hydrogels

To evaluate the osteogenic differentiation behavior of MC3T3-E1 cells on the PEG hydrogel surface, we constructed a cell culture model using cell-adhesive PEG hydrogels. Because of their nonadhesive nature, conventional PEG hydrogels do not exhibit favorable cell adhesion behaviors [10]. Consistent with this, both the 4-arm and 8-arm PEG hydrogels exhibited limited cell adhesion capacities. On unmodified PEG hydrogels, slight cell adhesion was observed on Days 1 and 3 of culture, with the behavior being more pronounced in 8-arm PEG hydrogels (Fig. 3a). By Day 5, the cells in both hydrogels had detached and formed spheroid clusters (Fig. 3a). These results further confirm the high bioinertness of the unmodified PEG hydrogel surface [37].

Fig. 3
figure 3

Cell adhesion and proliferation on the surface of hydrogels. a MC3T3-E1 cell growth on 0, 0.2, and 0.4 mM RGD-modified PEG hydrogels. The cells cultured without hydrogel were used as the control. Scale bars, 100 µm. b MC3T3-E1 cell proliferation by CCK-8 assay on 0.2 and 0.4 mM RGD-modified PEG hydrogels. The cells cultured without hydrogel were used as the control. Statistical analysis was performed via two-way ANOVA and Tukey’s multiple comparisons test. *P < 0.05, ns not significant

Therefore, the immobilization of cell-adhesive molecules is essential for evaluating cell differentiation behavior. Our goal was to establish a cell-adhesive surface by modifying the terminus of PEG-MA with the cell-adhesive RGD peptide. On the RGD-modified PEG hydrogel surfaces, both types of hydrogels exhibited good cell adhesion; however, the 8-arm PEG hydrogel demonstrated superior adhesion (Fig. 3a). The adhesion morphology and distribution of cells on the surface of the material were closer to those of the Petri dish without the hydrogel group with higher RGD concentrations (Fig. 3a). The results of the cell viability assay suggested that higher RGD modification rates significantly increased the viability of both the 4-arm and 8-arm PEG hydrogels (Fig. 3b). Although the concentration of RGD has been reported to vary between 0 and 5 mM [38,39,40,41], the choice of concentration is largely related to the type of substrate hydrogel employed and the cell type used in different scaffolds. In this study, the use of 0.2 mM and 0.4 mM RGD achieved cell adhesion. Immediately following this, the elastic modulus of the RGD-modified PEG hydrogels showed that, as the modification concentration increased, the storage modulus G’ decreased slightly. However, the elastic modulus of 8-arm PEG trended less than that of the 4-arm PEG hydrogels at the same concentration of RGD (Supplementary Fig. Sa). As RGD is terminated by a thiol group (RGD-SH), a Michael addition reaction occurs with PEG-MA. This affected the connectivity between PEG-SH and PEG-MA, leading to a decrease in the storage modulus [42]. Therefore, the inference is that RGD modification has an upper limit in this hydrogel system. After RGD modification, the swelling behaviors of both hydrogels remained consistent with those before modification (Supplementary Fig. Sb), indicating that RGD had little effect on the physical properties of the hydrogels. Based on the consideration of the impact on physical properties and modification effects, we used 0.2 mM and 0.4 mM concentrations to establish a cell culture model and evaluate the cell differentiation behavior on the hydrogel surface.

Osteogenic differentiation in RGD-PEG hydrogels

To test osteogenic differentiation, we subsequently evaluated the expression of Alpl and Ibsp, which are osteogenic differentiation marker genes, on Days 3 and 7. RT‒qPCR analysis revealed that the 8-arm PEG hydrogel modified with RGD presented greater gene expression than the 4-arm PEG hydrogel modified with RGD on Day 3 (Fig. 4a). Furthermore, the expression of osteogenic markers by the 8-arm PEG hydrogel modified with 0.4 mM RGD was still significantly greater than that by the 4-arm PEG hydrogel modified with 0.4 mM RGD on Day 7 (Fig. 4b). These results suggest that the increased phase separation in the 8-arm PEG hydrogel enhanced osteoblast differentiation.

Fig. 4
figure 4

Osteogenic induction of MC3T3-E1 cells. Relative mRNA expression of Alpl and Ibsp at 3 days (a) and at 7 days (b) in the control (Ctrl) without hydrogel and 0.2 and 0.4 mM RGD-modified PEG hydrogels. The cells cultured without osteogenic induction were used as the negative control (Neg) for each group. The data are presented as the mean ± SD, and n = 3 in each group. Statistical analysis was performed via Student’s t test. *P < 0.05, **P < 0.005, ns not significant

The mechanisms underlying the increased osteogenic induction by 8-arm PEG hydrogels with RGD modifications remain to be understood. Generally, osteogenic differentiation has been reported to be more effective on hard surfaces [43, 44]; however, the 8-arm PEG hydrogel was softer than the 4-arm PEG hydrogel was, suggesting the involvement of other mechanisms. The surface structure of materials, such as surface chemistry, mechanical properties and topography, can influence various cellular events, including cell morphology, mechanotransduction, and later tissue development [45,46,47]. In this study, the surface characteristics of 8-arm gels and 4-arm gels were notably different, which is hypothesized to affect the behavior of MC3T3-E1 cells. However, further experiments are needed to evaluate the relationship between specific cell behaviors and material surface structures to confirm these effects.

In summary, this study enabled a simple analysis of osteoblast differentiation behavior on phase-separated hydrogel surfaces. Its effectiveness surpassed that of standard tissue culture-treated polystyrene (TCPS) plates, making it a potentially efficient scaffolding material for future osteogenic differentiation.

Conclusion

This study demonstrated that 8-arm PEG hydrogels achieving GGPS are highly effective for osteogenic applications. The key findings include enhanced hydrophobic behavior, along with superior cell adhesion and differentiation, compared with those of 4-arm PEG hydrogels and traditional TCPS plates. The study found that these hydrogels, especially the 8-arm variant, not only presented increased particle adsorption owing to their hydrophobic surfaces but also significantly promoted osteogenic differentiation. This is attributed to the unique phase-separation properties of the hydrogel, which are crucial for enhancing its biological efficacy. Therefore, 8-arm PEG hydrogels represent a promising and innovative material for osteogenic tissue engineering and offer a simple yet effective solution for advancing regenerative medicine.