Study on galactosylated sodium alginate for enhancing HepG2 Cells adhesion and 3D printability

Abstract Sodium alginate is a polyanionic natural polysaccharide polymer widely used in tissue engineering. However, the lack of binding domains for interaction with cells limits its application in regenerative medicine. This study designed a kind of galactosylated sodium alginate (G-SA) material with improved galactose grafting rate by EDC/NHS activation of carboxyl groups in MES buffer and subsequently cross-linking by Ca2+ aims to enhance the adherence behavior of HepG2 cells on alginate substrate. The synthesized G-SA was characterized by Fourier transform infrared spectra and nuclear magnetic resonance spectroscopy. G-SA exhibited good biocompatibility and significantly enhanced the adhesion behavior of HepG2 cells on its surface. Furthermore, we demonstrated that the effect of G-SA concentration in enhancing cell adhesion was diminished at higher than 2% w/v. Finally, the suitability of G-SA material is investigated for 3D printing, demonstrating that HepG2 cells could maintain high viability and excellent printability in the interior of the gel. In addition, cells could multiply and grow into cell spheroids with an average size of 200 μm in G-SA scaffolds. These results indicated that galactosylated sodium alginate material could be used as a 3D culture system that could be effective for engineering liver cancer models.


Introduction
Natural polymers are hydrogel materials that play an essential role in simulating extracellular matrix in vitro [1][2][3][4]. However, polymers like gelatin [5,6], collagen [7,8], and fibronectin [9,10] usually possess binding domains that promote cell adhesion. At the same time, polysaccharides (starch, pectin, and alginate) tend to exhibit poor cell adhesion [11][12][13][14]. Sodium alginate (SA) is one of the most popular natural polysaccharides, consisting of copolymers of 1,4-b-D-mannuronic acid and a-l-guluronic acid and composed of different ratios of GM, MM, and GG fragments. Due to their unique functionalities and properties [15,16], such as excellent biocompatibility, simple cross-linking, and biodegradability, they have been widely used as bioinks [17,18], wound dressings [19,20], and tissue engineering scaffolds [21,22]. However, SA lacks a binding domain for cell interactions and polyanionic properties. Hydrophilicity tends to form a hydrated layer on the surface, resulting in poor cell adhesion and growth on its surface, hindering its application in the biomedicine fields [23,24].
Therefore, new methods have been developed to enhance the bioactivity of sodium alginate to enhance the interactions with cells and to expand its application in tissue engineering and the construction of in vitro models. For instance, by investigating the adhesion effects of different molecular structural compositions of sodium alginate on human osteoblast cell lines, mouse fibroblasts, and human keratin-forming cells, Dodero and colleagues [25] demonstrated that polysaccharides with an evident polyelectrolyte nature were found to better promote osteoblast cell viability and adhesion compared to the ones displaying a nearly neutral behavior. However, it is relatively challenging to facilitate the implementation of adhesion of different cells through the regulation of the molecular structure itself. Therefore, it has become a universal approach by compounding sodium alginate with other bioactive materials (gelatin, collagen) that have cell adhesion sites [26][27][28]. Nevertheless, these materials often fail to achieve cell-specific adhesion and are expensive. With further investigation of cell adhesion behavior in recent years, the introduction of cell adhesion molecules (CAM) or bio-responsive materials (RGD peptides and other cell adhesive peptides) into the matrix can effectively enhance cell adhesion and growth in the matrix [29][30][31]. For instance, Jeong et al. [32] developed a cell adhesive peptide (GRGDSP)-modified alginate (RA) electrospun nanofiber that vastly improved the ability of human skin fibroblasts to adhere, spread, and proliferate.
Galactose is a specific ligand for the salivary glycoprotein receptor (ASGPR) on the surface of hepatocytes, and ASGPR and glycosylated ligands can mediate cell adhesion to maintain a stereoscopic shape similar to that in vivo and maintain differentiation over time. Therefore, the design of galactose matrix materials is one of the hot spots for studying matrix materials for in vitro culture of hepatocytes [33]. In this study, galactose-modified sodium alginate hydrogels are fabricated for biomedical applications. The conveniently grafted galactose into sodium alginate in the aqueous phase using EDC/NHS reaction formed gels using CaCl 2 crosslinking, which conferred good biocompatibility and cell adhesion to galactose sodium alginate hydrogels. In addition, we have explored the printability of G-SA, using 3D printing technology to construct a cell-laden scaffold. Our results demonstrate that galactosylated sodium alginate could effectively enhance the adhesive growth of HepG2 cells in a suitable concentration range. Moreover, the fabricated bioink with biocompatibility and printability facilitated the aggregated growth of HepG2 cells in 3D conditions. We hope that our exploration of G-SA will enable its further application in tissue engineering scaffolds and in vitro liver model construction (Scheme 1).

Synthesis of galactosylated sodium alginate (G-SA)
0.15 g of sodium alginate was dissolved in 40 mL of MES (0.2 M, pH 4.5) buffer and stirred for 3 h until completely dissolved. Then, 17.5 mg of NHS and 363.05 mg of EDC (molar ratio EDC: COO-¼ 2.5, NHS: EDC¼ 0.2). NHS: EDC ¼ 0.2) were added and stirred for 30 min at room temperature until homogeneous. Further, 245.1 mg of galactosamine hydrochloride (molar ratio Gal:monosaccharide ¼ 1.5) was added and reacted for 24 h. The reaction product was placed in a cellulose dialysis bag with MWCO of 10,000 cellulose dialysis bag dialyzed with deionized water for 3 days, and then lyophilization. Scheme 1. Representation diagram of G-SA synthesis and application, synthesis of G-SA, promotion of HepG2 cell adhesion, and 3D bioprinting of G-SA bioink.

Structural characterization of G-SA
FTIR analysis: A certain amount of sample was mixed with potassium bromide, ground, and pressed using the potassium bromide press method. The FTIR characterization was performed on a Nicolet/Nexus 670 FTIR Analyzer (Nicolet Co., USA) with 64 scans in the range of 400-4000 cm À1 and a resolution of 4 cm À1 .
1 H NMR test: G-SA was dissolved in heavy water at a sample concentration of 5 mg/mL and temperature of 60 C. The sample structure was characterized by 1 H NMR using a 400 M superconducting Fourier Transform Nuclear Magnetic (Bruker AVANCE 400) spectrometer.

Thermal gravimetric and differential scanning calorimetry analysis
Thermal gravimetric analysis: the thermal resistance of G-SA samples ($ 5 mg) was examined using a TGA 50H in open pans under an argon atmosphere at a gas flow rate of 50 mL/min and a heating rate of 10 C/min from 35 to 500 C.
Differential scanning calorimetry analysis: to make the data more accurate, we conducted a thermal history elimination experiment on the material. The differential scanning calorimetry (DSC) measurement of G-SA samples ($ 5 mg) was carried out using a Discovery DSC 200 under a dynamic inert argon atmosphere, with a flow rate of 50 mL/min in a Tzero hermetic aluminum capsule. Heating was performed from 40 C to 300 C at a rate of 10 C/min.

Cytotoxicity of the G-SA extract and precursors solution
MC3T3-E1 cells were cultured in DMEM containing 4.5 g/L of glucose, supplemented with 10% FBS, 2 mM of L-Glutamine, 100 units/mL of penicillin, and 100 lg/mL of streptomycin. Cells were maintained at 37 C, 100% humidity, and 5% CO 2 and were passaged every 2 days. The cytotoxicity of the hydrogels was performed by two methods, including the hydrogel extract assay and the direct contact between hydrogel and cells assay.
For the hydrogel extract assay, all hydrogel extracts with different concentrations (5,10,20,30,40, and 50 mg/mL) were obtained via incubation of hydrogels in DMEM media at 37 C for 24 h. MC3T3-E1 cells suspension was added to 96-well plates and cultured in an incubator with 5% CO 2 at 37 C for 24 h. Then, the cultured media were removed, the hydrogel extracting solution was added, and the solution was cultured for another 24 h. CCK-8 assay was used to determine the cell viability, and the cells incubated in culture media without the extracting solution were set as the control group. The cell viability (%) is shown as the percentage of the absorbance value relative to the control.
For the hydrogel precursors assay, a series of G-SA solutions (0.1, 0.5, 1, 3, 5, and 10 mg/mL) were obtained with culture media. MC3T3-E1 cells at an appropriate cell density were added to 96-well plates with 100 lL of media and cultured in an incubator with 5% CO 2 at 37 C for 24 h. Then, the cells were cultured in 100 lL of the above G-SA diluted solutions for another 24 h. CCK-8 assays were used to determine the cell viability, as described in the Section hydrogel extract assay.
2.6. HepG2 cells culture and cell adhesion on G-SA HepG2 cells were cultured in DMEM containing 4.5 g/L of glucose, supplemented with 10% FBS, 2 mM of L-Glutamine, 100 units/mL of penicillin, and 100 lg/mL of streptomycin. Cells were maintained at 37 C, 100% humidity, and 5% CO 2 and were passaged every 2 days.
We used the CCK-8 method to assess the proliferation of HepG2 cells. Briefly, 100 lL of G-SA gel solution (0.1%, 0.5%, 1%, 2%, 3%, and 4% (w/v)) were coated at the bottom of the 24-well plate and cross-linked by Ca 2þ . Next, 1 Â 10 5 cells were seeded onto the gel and incubated for an additional time (1, 3, and 5 days). The medium was then removed, followed by adding 50 lL of CCK-8 reagent and 450 lL DMEM to each well and incubating for 2 h at 37 C. The absorbance was measured using a microplate reader at 450 nm. The effect of the G-SA gel on HepG2 cells adhesion was determined using fluorescence imaging of the AO/EB staining method, and the number of cells was calculated with ImageJ software. In addition, cytoskeleton staining was performed using actin-tracker red rhodamine to observe the morphology of HepG2 cells on the surface of G-SA gels at different concentrations.

Preparation and characterization of G-SA bioink
In this study, 2% (w/v) G-SA gel solution was chosen according to the results of section 2.6 by mixing 8% (w/v) gelatin solution in equal volumes to form the bioink at the G-SA and gelatin concentrations of 1% and 4%, respectively. The details are as follows, the precursor solution with the concentration of 4% (w/v) and 8% (w/v) was prepared by dissolving G-SA and gelatin in ultra-pure water. Then, the same volume of G-SA and gelatin precursor solution was mixed, and the vortex mixing was carried out. Transfer the mixed solution to the refrigerator at 4 C for physical cross-linking, then take it out and add 100 mM CaCl 2 solution for complete cross-linking. The fabricated G-SA/gelatin composite hydrogel was characterized as follows.
Rheological characterization: The rheological properties of different solutions were analyzed using a rotational rheometer (Anton-Paar, MCR302, Graz, Austria). The viscosity values of G-SA/gelatin hydrogels at various concentrations were measured by changing the shear rate from 0.01 to 100 S À1 . Then, the storage (G') and loss (G'') modulus of G-SA/gelatin were evaluated by changing the temperature from 40 to 5 C (cooling rate: 5 C/min.), Swelling behavior: The swelling tests were carried out in ultrapure water for various composite hydrogel samples. First, the cross-linked G-SA/gelatin composite hydrogel was freeze-dried, weighed the mass was noted as Wo. Then, the composite hydrogels were placed in water for 12 h. The masses at 1, 2, 3, 4, 5, 6, and 12 h were weighed and recorded as Wt. The swelling ratio (SR) was calculated according to the following equation: Degradation behavior: To test the degradation performance, firstly, the G-SA/gelatin solution was cross-linked by CaCl 2 solution to make a cylindrical sample and the mass was weighed and recorded as Wi. After swelling equilibrium, the samples were immersed in a DMEM medium and placed in a 37 C cell incubator. After 3, 6, and 9 days, the samples were removed using absorbent paper to wipe off the residual liquid on the surface, and the mass was recorded as Wc. The hydrogel degradation rate was calculated as: Morphology observation: The morphologies of the G-SA hydrogel samples were observed by scanning electron microscopy (SEM). The hydrogels were immersed in deionized water to achieve swelling equilibrium, and swelling equilibrium hydrogels were transferred into liquid nitrogen to freeze and retain the swelling state. Then, swollen hydrogels were freeze-dried in a freeze dryer under a vacuum at À50 C for 48 h. The dried samples were attached to the sample table and coated with gold for 30 s for SEM observation. The interior microstructures of hydrogels were analyzed by SEM.

3D Bioprinting
Briefly, HepG2 cells were collected by centrifugation at 1000 r min À1 for 4 min and suspended in the G-SA and gelatin mixed solution to a density of 1 Â 10 6 cells mL À1 . Then, we placed the prepared printing bioink in a 4 C refrigerator for 10 min for reaching the pre-gel state to print as soon as possible to keep the cell viability at a high level. In addition, a 3D bioprinting system (Bio-ArchitectV R Pro, Regenovo, Hangzhou, China) was applied to fabricate 3D scaffolds with low pressure (0.2 bar) and fast printing speed (20 mm/s) and temperature control (Nozzle temperature: 25 C, Base temperature: 4 C). After fabrication, 100 mM CaCl 2 solution was used to crosslink the printed structure for 10 min. Finally, the printed scaffold was transferred to a 6-well plate, rinsed with PBS, and incubated in an incubator with a culture medium.

Cell viability, proliferation, morphology assessments and liver functional assay
The cell viability assay was performed on printed cell-laden hydrogel using the Live/Dead Kit and fluorescence microscopy. At each time point (day 1, 3, 5), the culture medium was removed from culture plate wells and the hydrogel was washed with DMEM and stained in 2 mM calcein-AM and 0.5 mM EthD-1 solution in DMEM for 30 min in a 37 C, 5% CO 2 incubator. The constructs were washed with DMEM twice and imaged using a laser confocal microscope (Leica TCS SP8, Germany). Albumin secretion and urea nitrogen synthesis were measured by albumin assay kit and urea nitrogen content assay kit.

Structural characterization
Galactosamine was reacted with SA in MES buffer to prepare galactosylated sodium alginate derivatives (G-SA, as shown in Figure S1, different environments of H have been marked). The 1 H NMR spectra of SA and G-SA are shown in Figure 1(a). It was observed that a new peak appeared between 3 and 4 ppm for G-SA, and a peak evident near 5.8 ppm could be attributed to the galactose [34]. The FTIR spectra of SA and G-SA are given in Figure 1(b). It could be observed that the G-SA resulted in peaks of 1608 cm À1 and 1411 cm À1 compared to SA, which could be ascribed to symmetric and antisymmetric stretching vibrations of the carboxylate ion -CO2-. However, these characteristic absorption peaks of carboxylate ions were significantly weaker compared with SA. In addition, the peak at 1542 cm À1 could be attributed to the 'acrylamide II peak' (C (C-N-H bending vibration), demonstrating that the galactose group was attached to the SA through the formation of an acyl group by the amino and carboxyl groups. Meanwhile, the acylamide II peak (C-N-H bending vibration) at 1542 cm À1 was enhanced, demonstrating that the galactose moiety was attached to SA through the formation of acyl groups by amino and carboxyl groups [35,36].

Thermal properties
TGA was conducted to explore the stability of G-SA (Figure 1(d)). About 20% (wt/wt) of weight loss was observed between 40 and 210 C, where the highest loss occurred below 110 C. Accordingly, a heat absorption peak appeared on the DSC curve (Figure 1(c)), which could be attributed to moisture loss. Between 210 and 500 C, a weight loss of nearly 40% (wt/wt) was observed, which could be due to the damage caused to the crystal zone portion of sodium alginate during the temperature rise, causing it to melt. The melting temperature of G-SA was lower than SA, mainly caused by the partial disruption of the SA crystal region due to the modification of galactose. Further, the rise in the temperature to above 250 C resulted in decomposition. This exothermic process with a sharp peak on the DSC curve showed consistency with the TG results.

Biological properties
The cell compatibility was evaluated by the G-SA direct contact test and extract assay. For the direct contact test, the cytotoxicity evaluation of G-SA precursors to MC3T3 was performed using CCK-8 assay at six different concentrations of G-SA in DMEM media over 72 h: 0.1, 0.5, 1, 3, 5, and 10 mg/mL, as shown in Figure 2(a) and Figure  S2. G-SA precursors showed no signs of cytotoxicity to MC3T3 within 48 h at low concentrations (0.1-5 mg/mL), while the cell survival rate of the 10 mg/mL group resulted in a decrease in cell viability to about 80%, which might be due to the inhibition of cell proliferation caused by the increased viscosity of the culture medium due to G-SA. After culturing up to 72 h, the cell survival levels in the 5 mg/mL (83.42% cell survival) and 10 mg/mL (72.99% cell survival) groups showed significant differences from the 0.1 mg/mL to 3 mg/mL (>95% cell survival) groups. Specifically, the treatment of the 10 mg/mL group indicated a cell survival rate of only about 70%, indicating that the relatively higher concentrations of G-SA precursors might have caused some cytotoxicity. In comparison, lower concentrations showed no side effects on cells.
As shown in Figure 2(b) and Figure S3, for the G-SA extract test, even though the cell survival rates of the 50 mg/ml and 5 mg/ml groups produced significant differences at incubation up to 72 h, the cell survival rates of the groups were still above 90%, indicating that the groups did not show apparent cytotoxicity to MC3T3 cells.
In addition, the live-dead staining assay also demonstrated the excellent cytocompatibility of the synthesized G-SA. Figure 2(c) and (d) shows the growth of MC3T3 cultured for 72 h in G-SA precursors and extracts, respectively. As shown in Figure 2(a), except for the concentration groups of 5 mg/mL and 10 mg/mL where more red dots (i.e. dead cells) appeared, there were only a very few dead cells in the other concentration groups, and the vast majority were green dots (i.e. live cells). Figure 2(d) indicates the results of live-dead fluorescence staining of MC3T3 cells in G-SA extract, and it can be found that the cell growth status of each group was good, and few dead cells appeared. The complete staining pictures of live and dead cells are shown in Figure S4 and Figure S5. The above results are consistent with the cell viability data shown in Figures 2 (a) and (b), which proves that G-SA has good cytocompatibility and has specific application prospects in the biomedical field.

HepG2 cell adhesion
We seeded HepG2 cells to study the cell adhesion capability of G-SA. The results are shown in Figure 3. Cells usually undergo initial cell attachment, spreading, organization of actin cytoskeleton, and focal adhesion formation to achieve stable adhesion on the substrate [36,37]. Sodium alginate is not adhesive to HepG2 cells because its molecular chain does not have a cell-binding domain. It is hydrophilic and therefore prevents cell adhesion by forming a hydrated layer on the surface. In addition, the negatively charged sodium alginate under physiological conditions would repel negatively charged cells [23,24]. Six different concentration groups of 0.1, 0.5, 1, 2, 3, and 4% (w/v) were set to investigate the adhesion of HepG2 cells on cross-linked G-SA substrates. On day 1, the cells in the 0.1, 0.5, and 1% (w/v) groups showed initial cell morphology. As seen in Figure 3(a), the cells in the 2% (w/v) group showed no apparent specific morphology, and the HepG2 cells in the 3% (w/v) and 4% (w/v) groups showed round shape, indicating that the effect of different concentrations of G-SA on cell morphology was different. This difference was more pronounced from the third day onwards and produced a disparity in cell proliferation behavior. Below 2% (w/v) concentration, the proliferative behavior was more pronounced, with cells forming tight focal adhesions on the G-SA gel matrix. On the contrary, the cells remained single, contrary to the growth behavior of HepG2 cells that grow in aggregates with tight junctions. Further, the number of cells per unit area was counted ( Figure S6. The number of cells per mm 2 increased from 90, 147, 125, 94, 52, and 53 to 558, 1071, 886, 192, 155, and 119, respectively. Similar results appeared on day 5, indicating that cross-linked G-SA substrates below 2% (w/v) concentration were more compatible with HepG2 cell growth and adhesion, especially 0.5 and 1% (w/v). The same conclusion was obtained from the CCK-8 assay ( Figure S6). In addition to comparing cell numbers, cytoskeleton staining was performed using phalloidin to observe the cell spreading and adhesion status of HepG2 cells on the surface of different concentrations of G-SA gels. In addition to comparing cell numbers, the cytoskeleton staining using phalloidin showed the cell spreading and adhesion status of HepG2 cells on the surface of different concentrations of G-SA gels. As shown in Figure 3(b), HepG2 cells were inoculated on the surface of G-SA gels at concentrations of 0.1, 0.5, 1, 2, 3, and 4% (w/v), respectively, and there were significant differences in cell spreading status. On the first day of inoculation, cells on the surface of G-SA gels with low concentrations (0.1, 0.5, 1% (w/v)) were already well spread, and filamentous pseudopods could be observed. On the contrary, the cells on the surface of G-SA gels with 2, 3, and 4% (w/v) concentrations showed a round shape, indicating that the cells might not have adhered well to the surface of the matrix. On days 3 and 5, in addition to the difference in the number of cells, the cell adhesion and growth status also differed significantly. The cells on the surface of the G-SA gel at low concentrations (0.1, 0.5, 1% (w/v)) showed obvious aggregated growth with clear and abundant morphology. In contrast, the cells at high concentrations (2, 3, 4% (w/ v)) also underwent proliferation but did not show the typical shape of HepG2 cells. Therefore, the effect of G-SA on cell adhesion was conditional, i.e. the concentration below 2% would have a pro-adhesive effect, while above 2% would result in poor cell growth and adhesion. Although some studies using the specific binding of galactose to ASGPR to enhance the adhesion behavior of hepatocytes or hepatocellular carcinoma cells have been reported, a detailed discussion of the effect of galactose concentration on cell adhesion growth behavior has not been made [38,39]. Based on the experimental results, it was evident that the inhibitory effect of G-SA gel on cell adhesion above 2% concentration might be due to the oxidative damage caused by galactose to HepG2 cells [40,41]. Therefore, it is significant to determine the appropriate concentration of G-SA for culturing HepG2 cells.

Characterization of G-SA/gelatin composite bioink
Since the rheological properties of hydrogels play an essential role in the printability and extrudability of 3D printing processes [42], the shear viscosity variation properties and the storage and loss modulus of G-SA hydrogel solutions were analyzed by an MCR 302 rheometer. As shown in Figure 4(a), the viscosity of hydrogel solution gradually decreased with an increase in the shear rate between 0-100 S-1, representing a type of non-Newtonian fluid properties. In addition, this property is one of the necessary conditions for extrusion 3D printing because bioink could be squeezed out of the nozzle with minimal pressure, effectively protecting cell viability and maintaining the fidelity of the printed structure [43]. Moreover, the shear viscosity of G-SA/gelatin and the viscosity change trend of SA/gelatin are roughly similar.
Nevertheless, the printing ink containing G-SA has a relatively lower viscosity, possibly due to the viscosity reduction caused by the formation of intramolecular hydrogen bonds (N-H-O) after galactose grafting modification. In addition, to determine the gel temperature of G-SA/gelatin composite hydrogel, we examined storage modulus and loss modulus with temperature, and the results are shown in Figure 4(b). Indeed, the bioink formation pre-gel is critical for extrusion 3D printing, possessing a significant impact on print formability and resolution. It was observed that both groups of printing bioinks could become gel around 20 C (G-SA/gelatin: 20.9 C, SA/gelatin: 17.5 C), and the boink containing G-SA resulted in a relatively higher gel point temperature. This phenomenon could be more conducive to achieving pre-gel conditions before printing, thus reducing the preparation time before printing and the damage caused to cells. Besides, as can be seen from the fitted results in Figure 4(a), the rheology data are described well by Carreau fluid model, the equation is Where l inf , l 0 , k and n are material coefficients, they represent viscosity at infinite shear rate (PaÁs), viscosity at zero shear rate (PaÁs), characteristic time (s) and power index (-), respectively. a (-) is transition control factor. The Carreau model treat the fluid as a Newtonian fluid at a high shear rate and as a Power law fluid at a low shear rate. According toe the value of the power index n, the fluids can be classified into: Shear-thinning fluids for which n is less than 1, Newtonian fluids for which n is 1, Shear-thickening fluids for which n is greater than 1 (Table 1). Therefore, both fluids are Shear-thinning fluids in our experiment, however, SA/Gelatin has a higher degree of shear-thinning behavior than G-SA/Gelatin. Further, the swelling properties of G-SA/gelatin composite hydrogel were evaluated in Figure 4(c). As shown in the Figure, both hydrogels rapidly absorb water and swell within a few hours, and then slowly reach swelling equilibrium. After 12 h, the dissolution rate of SA/gelatin hydrogel is greater, reaching 6.92 times, while that of G-SA/gelatin is 6.05 times. Both hydrogels are hydrophilic hydrogels and the swelling kinetics exhibit a controlled process of diffusion, i.e. the diffusion of water molecules into the gel [44]. the internal pores of G-SA are larger and water molecules enter more quickly to reach swelling equilibrium, a result that is consistent with the examination of the internal morphology of the hydrogels.
The degradation property of G-SA/gelatin composite hydrogel was evaluated in Figure 4(d). From the in vitro degradation curves, it can be seen that the degradation rate of both hydrogels was faster in three days and started to slow down later, which is because gelatin is a polymeric material with high protein content and degrades rapidly in the culture medium [45], in addition to the 37 C incubation conditions that accelerate the detachment of gelatin from the gel system. We examined the degradation of G-SA and SA as well as G-SA/gelatin and SA/gelatin hydrogels separately within the seven-day degradation examination period and found that none of the degradation rates exhibited significant differences. Furthermore, to investigate the internal morphology, the cross-sectional morphologies of G-SA/gelatin and SA/gelatin hydrogels were characterized by SEM (SEM, S-4800, Hitachi, Tokyo, Japan) as shown in Figure 5(a) and (b). By observing the internal structures of the hydrogels after lyophilization, it could be found that both hydrogels did not exhibit significant phase separation, which indicated that SA and G-SA could be well mixed with gelatin. In addition, the hydrogels exhibited a threedimensional porous structure similar to that of natural porous hydrogels, and the hydrogels had a similar pore size shown in Figure S7 (G-SA/gelatin: 193.84 um, SA/ gelatin: 169.86 um), and the pore size of the G-SA/gelatin composite hydrogel was slightly larger.

3D Bioprinting of G-SA/gelatin composite scaffold
Considering the rheological properties, the prepared bioink consisting of G-SA/gelatin encapsulated HepG2 cells is suitable for extrusion-based 3D bioprinting by adjusting parameters, including printing speed, pressure, and temperature. SA/gelatin is already a mature bioink with excellent printability [46,47]. As depicted in Figures  5(c) and (d), both bioinks could be well molded, resulting in an intact and stable scaffold structure after cross-linking. SEM images showed a uniform size (pore size around 500 lm) and precise contour of the printed grid structure, appearing white after lyophilization.

Bioprinting and cell viability assessments of HepG2 cell-laden 3D scaffolds
3D bioprinting has garnered enormous interest from researchers in constructing tissue engineering scaffolds due to its facile construction of bulk constructs [48]. In this research, HepG2 cells wrapped in G-SA/gelatin ink were printed as 3D scaffolds and cultured for 1, 3, and 5 days. Then, cell proliferation and growth status of cells were examined using the CCK-8 and live-dead cell staining kits. An apparent green fluorescence was visible on the G-SA/gelatin scaffold, indicating that HepG2 cells showed excellent cell viability on the scaffold, mainly due to the efficient nutrient transport under 3D culture conditions [49]. As shown in Figure 6(a), HepG2 cells in the scaffold containing G-SA showed significant cell aggregation on the third day, while the control group showed only a minimal number of cell spheroids. The improved growth of HepG2 cell aggregation could be accelerated in the presence of galactose in the culture medium, mediated by the action of ASGPR with glycosyl ligands, and the number of cell spheroids by the fifth day held. This expected result could be applied in the in vitro construction of liver-specific tumor models [50][51][52]. As shown in Figure 6(b), cell proliferation results showed that cells maintained continuous proliferation in the scaffold. In contrast, the cell proliferation level in the scaffold containing G-SA was higher than that in the control group, producing a significant difference on the fifth day, mainly due to the promotion effect of G-SA on HepG2 cells. Meanwhile, the collected cell culture supernatants were continuously collected on days 1, 3, and 5 of cell culture and measured the albumin and urea nitrogen secretion of the cells. As shown in Figure 6(c) and (d), it was evident that the secretion increased with increasing culture time, and the secretion was greater in the G-SA group compared to the SA group. It was demonstrated that the cells grew and proliferated well in the 3D scaffold. The G-SA group showed higher hepatic functional protein secretion and urea synthesis, indicating that the formed 3D spheroids possessed better hepatic functional activity. Our findings could expand novel bioink materials for 3D bioprinting technology to build liver cancer models or liver microtissues, which presented excellent interaction with HepG2 cells towards potential biomedical applications.

Conclusions
In this study, a sodium alginate derivative grafted from galactose was developed, and its effect on HepG2 cell adhesion and 3D printability was investigated. Compared to many materials with cell adhesion-promoting properties (e.g. gelatin, decellularized matrices, and RGD peptides), sodium alginate possesses minimal cell adhesion-promoting properties, and even its polyanionic properties inhibit cell stretch adhesion. In this study, G-SA was prepared to overcome the poor cell binding domains of SA, and had good printability. More importantly, compared with above materials, such as gelatin, dECM, and RGD peptide, SA-based materials have better crosslinking and machinability, which has greater application potential in micro-nano processing and bio-manufacturing. The results showed G-SA could specifically promote cell adhesion to HepG2, which is not available to other materials and is also one of the important innovations of this study. Besides, although the use of galactose-grafted natural macromolecules to promote hepatocyte adhesion has been reported in the literature, SA, as a highly used biomaterial, has not been adopted this strategy for 3D bioprinting to utilize it deeply in the construction of tissue engineering and drug screening models. We investigated the concentration of G-SA suitable for HepG2 cell adhesion. Bioink with good printability was obtained by composite gelatin, and 3D scaffolds encapsulated with HepG2 were successfully printed, which well supported the growth and proliferation of cells under 3D culture conditions. Surprisingly, HepG2 has apparent aggregation growth behavior in the G-SA/gelatin scaffold, which is expected to be further applied to the study of tumor cell morphology and high throughput screening of anti-cancer drugs under in vitro 3D culture conditions.