Collagen I is a key extracellular matrix (ECM) component in bone tissue and one of the most important biomaterials for bone tissue engineering applications. However, printing high-resolution mesh scaffold from collagen I remains challenging due to its relatively weak ink shape fidelity. While previous efforts have attempted to improve printability by increasing ink viscosity, such approaches often compromise ink flowability and yield only modest improvements in printing resolution. To solve this issue, we blended oxidized cellulose with collagen I to form a Schiff-base interaction. The resulting hydrogel exhibited lower viscosity but a more apparent linear rheological characteristic, as demonstrated by our large amplitude oscillation sweep results. This enhanced rheological profile enabled the fabrication of scaffolds with a printing resolution approaching 150 μm—one of the highest reported for collagen I-based scaffolds. Scaffolds with this scale of rod diameter and pore size greatly enhanced the proliferation and osteogenic differentiation of mesenchymal stem cells. Correspondingly, the expression of key osteogenic markers, including N-cadherin, HIF-1α, and β-catenin, was upregulated. These findings broaden our understanding of scaffold design and processing optimization of collagen I-based scaffolds and may advance their application in bone tissue engineering.

1. Introduction

Collagen I is a key extracellular matrix (ECM) compound in bone tissue and has been considered one of the most ideal biomaterials for bone tissue engineering.^1,2 It is highly biocompatible and contains surface Arg-Gly-Asp (RGD) groups that bind to specific cell receptors, promoting cell adhesion.^3,4 Therefore, it is superior to general thermoplastic biopolymers and ceramics for a broad field of tissue engineering applications. Also, collagen I has been shown to promote osteogenic differentiation of stem cells more effectively than many other natural biopolymers like gelatin.⁵ For these reasons, some studies have blended it with other biopolymers or ceramics to produce bone tissue scaffolds.^6–9 However, printing a high-resolution scaffold composed of collagen I is challenging, presumably due to its weak gelation strength, especially in comparison to gelatin.^10,11 While gelatin macromolecules undergo a unique transition from random flexible coils to rigid triple helix structures under cooling, this process does not occur in collagen I.¹² Instead, its viscosity increases only through greater chain entanglement at lower temperatures.¹³ As a result, the shape fidelity of collagen I-based hydrogels is relatively weak, yielding generally poor printing resolution. High scaffold resolution is crucial to provide a large surface area for cell adhesion, absorption of nutrition and growth factors, and biochemical reactions necessary for ECM synthesis,^14,15 as well as to generate a high number of rod junctions to facilitate cell migration according to the Random Walk Model.¹⁶ Therefore, it is important to achieve high resolution with collagen I-based scaffolds so that both structural and compositional characteristics can be optimized simultaneously.

To achieve this goal, recent studies have utilized cryogenic printing to facilitate ink solidification. In this process, the hydrogel ink is printed onto a readily-cooled 2D plate or into a 3D well.^6,17–20 To further improve the printing performance, some studies attempted to use lower temperatures or added rigid nanoscale fillers to maximize viscosity.^21–23 Although these approaches improve the shape fidelity and printing resolution to some extent, achieving a resolution finer than 200 μm remains difficult.^6,17–20 Also, increasing ink viscosity inevitably reduces flowability, complicating extrusion and reducing the continuity and uniformity of the printed structures.^24,25

Given the low strength of the hydrogel formed by collagen I macromolecular entanglement, a promising solution is to introduce external bonding for stabilization. Among the commonly applied chemical modifications for hydrogel rheology, Schiff-base interactions are frequently used. This reversible bonding can be weakened by shear stress during extrusion and reformed afterward. So, Shiff-base hydrogels exhibit shear-thinning behavior, maintaining both flowability and shape fidelity.²⁶ Some recent studies have applied Shiff-base interactions to gelatin-based hydrogels, resulting in increased ink rigidity and mechanical properties.^27–30 Following this principle, we introduced Schiff-base interactions into a collagen I-based hydrogel to study its effects on rheological properties, printing performances, and its ability to promote mesenchymal stem cell (MSC) proliferation and osteogenic differentiation—assessing its potential prospects in bone tissue engineering.

In this study, we overcame the hurdle of printing high-resolution natural hydrogel-based mesh scaffolds by blending oxidized cellulose and collagen I to form a hydrogel ink with Schiff-base interactions and printing it under cryogenic conditions. The resulting scaffolds featured pore and rod sizes as small as 150 μm. We investigated how these dimensions influenced MSC proliferation and osteogenic differentiation, and discussed the possible underlying mechanisms. This study not only advances the understanding of optimal scaffold dimensions for osteogenic differentiation but also achieves simultaneous optimization of scaffold composition and structure. We believe this work contributes significantly to scaffold design and optimization for bone tissue engineering applications.

2. Materials and methods

2.1. Materials

Collagen I, alginate (β-D-mannuronicacid/-L-guluronicacid ratio (M/G) = 1:1), gelatin (gel strength ~200 Bloom, biotech grade), sodium peroxide, genipin, and anhydrous calcium chloride (CaCl₂) were all supplied by Macklin (China). Hydroxymethyl cellulose was provided by Sigma Aldrich (USA).

2.2. Oxidation of cellulose and preparation of hydrogel inks

Hydroxymethyl cellulose was dissolved in deionized water in a 4% w/v concentration (4 g in every 100 mL of water) in a 100 mL flask. The solution was heated at 37°С for 1 h and stirred constantly with a magnetic bar. Then, 0.2 w/v sodium peroxide was added to the solution in a dark environment, and the solution was stirred at 40°С for 1 h. The resulting solution was denoted as “oxidized cellulose solution”. A schematic of the reactions is shown in Figures 1 and 2. Then, 8% w/v collagen I and 8% w/v gelatin was added to the oxidized cellulose solution by stirring at 37°С for 1 h. Lastly, 25% w/v alginate was added to each solution and stirred at 37°С for 1 h. For the collagen I-based hydrogel inks, the ink containing non-oxidized cellulose was denoted as “ink A,” while the ink containing cellulose oxidized by sodium peroxide was denoted as “ink B”. In the meantime, a crosslinking solution was prepared by dissolving 2% w/v genipin and 20% w/v CaCl₂ in deionized water. Fourier transform infrared spectroscopy (FTIR) characterization was performed on hydrogels containing ink A or B using a Bruker FTIR instrument (Germany).

2.3. 3D printing and post-treatment handling

3D printing was conducted with an extrusion-based 3D printer (Regenovo, China). The printer was equipped with a 3D cryogenic well, and the extrusion force was driven by compressed air. Tapered plastic needles were used for printing (inner diameter: 200 μm; outer diameter: 220 μm). The temperature of the 3D cryogenic well was set to −1°С for printing both collagen I and gelatin-based inks, and a total of seven layers were printed with each ink. Three samples were printed using ink B and with rod distances (distance between the centerline of adjacent rods) of 300, 450, and 600 μm, and were denoted as “sample 1,” “sample 2,” and “sample 3”, respectively. Another scaffold was printed with gelatin-based ink and a rod distance of 600 μm, denoted as “sample 4”. The layer height and extrusion pressure were both set to 0.14 mm. The printing pressure was set to 0.18 MPa for printing collagen I-based ink and 0.25 MPa for printing gelatin-based ink. Upon the completion of printing, the crosslinking solution was added dropwise to the scaffolds until fully immersed. The scaffolds were then removed and stored in a refrigerator at −20°С.

2.4. Morphological study

All printed scaffolds were freeze-dried and loaded into the chamber of a scanning electron microscope (SEM; Zeiss, Germany) and vacuumed again. SEM observation was conducted under a secondary electron (SE) mode with an accelerating voltage of 3kV. ImageJ software (1.8.0, National Institute of Health, USA) was used to analyze the average rod diameter and pore size of all scaffolds.

2.5. Rheology

Rheological studies were conducted with an Anton Paar MCR 302e rheometer (Anton Paar, Austria), equipped with a 25 mm parallel plate and an isothermal chamber. Inks A and B were tested to evaluate the effect of Schiff-base formation on improving ink shape fidelity. Steady-shear flow test was performed from 0.1 to 100 s⁻¹ at −1°С for each ink. An oscillation temperature sweep test was conducted from 37 to −5°С at a constant frequency of 1 rad/s and a strain of 0.5%. Lastly, a large amplitude oscillation sweep (LAOS) test was conducted at a constant frequency of 1 rad/s across a strain range of 0.1–1000%.

2.6. Thermal study

Differential scanning calorimetry (DSC) test was conducted on hydrogel ink A and B using a HITACHI DSC 200 instrument (HITACHI, Japan) under a nitrogen environment. A cooling process from 30 to −10°С was applied at a rate of 1°С/min.

2.7. Mechanical study

Tensile tests were conducted on scaffold samples 1–4 using a universal test machine (UTM 4103, Shenzhen SUNS Technology Stock Co., Ltd., China). A loading rate of 1 mm/min was applied, and each group of scaffolds was tested in triplicate.

2.8. Cell culture and characterizations

2.8.1. Cell seeding

MSCs were extracted from 2-week-old Sprague Dawley mice. Before seeding cells, all scaffolds (samples 1–4) were immersed in 75% ethanol for 10 min, washed once with phosphate-buffered saline solution (Beyotime, China), exposed to UV radiation for 30 min, and immersed in an (α-MEM) Minimum Essential Medium (Vivacell, China) containing 10% fetal bovine serum (Gibco, Australia), penicillin (10 kU/mL) and 10 mg/mL streptomycin solution (10 mg/mL, Beyotime, China) for 24 h. This modified medium was denoted as “complete medium”. MSCs (passage 2) were seeded onto the scaffolds at a density of 10⁶ cells/mL. Cell culture was conducted in an incubator (Thermo Fisher Scientific, USA) at 37°С with 5% CO₂. After 24 h, once cell adhesion was observed, the complete medium was replaced with an osteogenic induction medium containing an additional 50 μM ascorbic acid, 10 mM β-glycerol phosphate, and 100 nM dexamethasone (all from Sigma Aldrich, USA).

2.8.2. Cell counting kit-8 test

On days 1, 3, and 5 of culture, a cell counting kit (CCK)-8 test was conducted to assess cell proliferation. The α-MEM medium containing 10% CCK-8 cell counting agent (Beyotime, China) was added and incubated for 1 h. A total of 100 μL of cell supernatant was used to measure the Optical Density (OD) value at a wavelength of 450 nm using a spectrophotometer (Multiskan GO, Thermofisher Scientific, USA).

2.8.3. Cell morphology staining

MSCs were seeded onto scaffolds placed in a 24-well plate. Actin-Tracker Red (Beyotime, China) and 4’6-diamidino-2-phenylindole (DAPI; Beyotime, China) were used for immunofluorescence staining to visualize the cytoskeleton and nuclei, respectively. An upright fluorescence microscope (DM6 B, Leica, Germany) was used for observation. The excitation/emission wavelengths were 496/516 nm for Actin-Tracker Red and 480/340 nm for DAPI, respectively.

2.8.4. Western blot

On days 7 and 21 of cell culture, protein was extracted from MSCs. Cells were lysed in Radio Immunoprecipitation Assay (RIPA) buffer (Beyotime, China), and total protein concentration was determined using the Bradford assay (Sangon Biotech, China). Protein of equal amounts was separated on a 12% Sodium Dodecyl Sulfate Polyacrylamide (SDS-PAGE) Gel Electrophoresis gel and transferred onto a 0.22 μm-pore size polyvinylidene difluoride membrane (Sigma-Aldrich, USA) under a constant current of 252 mA. The membranes were then blocked with 5% skim milk and incubated overnight at 4°С with primary antibodies targeting N-cadherin (NCAD), alkaline phosphatase (ALP), osteocalcin (OCN), hypoxia-inducible factor 1α (HIF-1α), β-catenin, and GADPH. After being washed three times with Tris-buffered saline with Tween-20 (Beyotime, China), the membranes were incubated at room temperature with horseradish peroxidase-conjugated secondary antibodies for 1.5 h. Lastly, an enhanced chemiluminescence detection kit (Beyotime, China) was used to visualize the immunoreactive bands.

2.9. Statistical analyses

Each experiment was performed independently in a minimum of three replicates. Statistical analyses were conducted using either Student’s t-test or one-way analysis of variance, as appropriate, to evaluate differences between treatment groups. All analyses were performed using GraphPad Prism 7 (GraphPad Software, Inc., USA). The p-value less than 0.05 was considered statistically significant.

3. Results

3.1. FTIR analysis

Following the oxidation of cellulose by sodium peroxide, the hydroxyl groups (–OH) were converted to aldehyde groups. In the FTIR spectra, ink A showed a characteristic –OH peak at 1075 cm⁻¹,³¹ which was markedly reduced in ink B, indicating successful oxidation. In addition, the characteristic amide peak of collagen I at 1542 cm⁻¹ was prominent in ink A but nearly absent in ink B (Figure 3), suggesting its consumption via Schiff-base reactions.³²

3.2. Rheological study

At the printing temperature of –1 °C, ink A exhibited higher viscosity than ink B. This was consistent with its lower tan δ value and steeper stress–strain curve slope, indicating greater stiffness (Figures 4 and 5). Both inks demonstrated shear-thinning behavior, as viscosity decreased monotonically with increasing shear rate (Figure 4B). LAOS analysis revealed more flattened Lissajous curves for ink B compared to ink A, indicating more linear viscoelastic behavior (Figure 6A). Additionally, ink B showed A higher strain-hardening ratio (Figure 6B), suggesting better structural recovery and printability. The greater linear rheological characteristic of ink B may contribute to its improved shape fidelity during printing.

3.3. Morphological study

Cryogenic 3D printing (Figure 7) was attempted using ink A and B. An optimum combination of pressure and speed parameters was evaluated for each ink until rod breakage occurred. Printing with ink A was more challenging due to its weaker gelation strength, which limited resolution improvements at higher printing speeds and led to frequent rod breakage (Figure 7B). In contrast, ink B enabled smoother printing, likely due to Schiff-base-induced structural reinforcement (Figure 7C). As a result, samples 1–3 scaffolds were printed using ink B, while sample 4 was printed using a gelatin-based hydrogel ink.

The microstructures of samples 1–4 were examined using SEM (Figure 8), and the average rod diameter and pore size were recorded (Table 1). The printing resolution approached 150 μm, representing one of the highest printing resolutions reported for collagen I-based scaffolds to date.³³

Table 1. Average rod diameter and pore size of scaffold samples before and after swelling

Note: Data are presented as mean ± standard deviations (n = 3).

To evaluate the dimensional stability during cell culture, microscale images of the scaffolds were also captured post-swelling (Figure 9). As shown in Table 1,only minor increases in rod diameter were observed, whereas the pore size exhibited slightly greater expansion. Overall, the swelling extent across samples 1–4 was significantly lower than typically reported values, which often exceed 100–200% dimensional increase.^34,35 This stability is likely attributable to the Schiff-base interactions between oxidized cellulose and collagen I.

3.4. Thermal study

The DSC curves for ink A and B are shown in Figure 10. Neither formulation exhibited an apparent phase transition process in the cooling process. Therefore, the printability and structural integrity of the scaffolds relied primarily on the intrinsic rheological properties of the hydrogel inks, including viscosity, viscoelasticity, and linear rheology. As discussed in Section 3.2., ink B demonstrated enhanced linear rheological characteristics of ink B—presumably due to Schiff-base interaction—which contributed to its superior shape fidelity.

3.4. Mechanical study

Tensile testing was conducted on scaffolds from samples 1–4, and the average elastic moduli were calculated from the linear regions of the stress–strain curves (Figure 11). The moduli were 11.41 ± 0.18 MPa (sample 1), 7.02 ± 1.35 MPa (sample 2), 6.99 ± 0.49 MPa (sample 3), and 4.75 ± 1.84 MPa (sample 4). The highest modulus in sample 1 can be attributed to its smallest rod diameter and pore size, which increased the number of crosslinking junctions and improved resistance to deformation under tensile stress.

3.5. Cell study

On day 5, the OD value of sample 1 was lower than both samples 2 and 3 (Figure 12), which was presumably caused by its lower porosity. According to the western blot (WB) results, the expression of OCN and ALP was the highest in sample 2, indicating the highest level of osteogenic differentiation of MSCs. The NCAD and HIF-1α expressions were also elevated in sample 2, reflecting a higher level of cell condensation and a more hypoxic condition. Lastly, the expression of β-catenin, an osteogenic-related marker, was also most upregulated in sample 2 (Figure 13).

4. Discussion

4.1. Ink rheological properties, printing quality, and post-treatments

Typically, the viscosity and rigidity of hydrogel inks, as measured through steady shear flow and oscillatory rheological tests, are used to evaluate shape fidelity during printing.²⁴ These are often conducted at a constant strain within the linear elastic region of the ink (0.5–1%).³⁶ When the hydrogel ink was printed at a relatively low speed, the induced strain remained within this linear range, making such tests appropriate for measuring shape fidelity. However, higher printing resolution often requires significantly higher speed, which can generate large strains—sometimes exceeding 100%.¹⁴ Under such large strain conditions, LAOS testing provides a more accurate reflection of the ink’s rheological behaviors.

Our results showed that although ink A demonstrated higher viscosity and stiffness under low-strain conditions (0.5%), it exhibited weaker linear rheological characteristics across the wider strain range (1–1000%). Linear rheology refers to the regime where stress is directly proportional to strain, resembling the behavior of an ideal Hookean spring.³⁷ In contrast, inks with strong non-linear behavior resemble viscous fluids and are less able to maintain internal stress under deformation.^38,39 Therefore, inks with stronger linear rheological behavior, such as ink B, are better suited for maintaining structural integrity during high-speed printing. This was further supported by the higher strain-hardening ratio observed in ink B.

The formation of the Schiff-base interaction may disturb the entanglement of collagen I macromolecules and reduce viscosity. Under a larger strain, however, the bonding strength of the Schiff base could be stronger than that of the collagen entanglement. Presumably, due to this reason, ink B with this Schiff-base interaction exhibited a stronger linear rheological characteristic and improved shape fidelity.

In this study, we utilized a moderate cryogenic printing condition at −1, which was unlikely to cause the denaturation of collagen I. According to previous studies, the collagen-to-gelatin transition must involve a heating process,^1,40,41 which was opposite to the temperature condition in cryogenic printing. There are also studies utilizing freeze-drying (at −20°C or lower) to produce collagen-based foam scaffolds, followed by cell and animal studies, indicating that the bioactivity of collagen composition was maintained.^42–45

For the crosslinking reaction, genipin was used in this study. The reaction involves the bonding of its aldehyde group to the amine group of collagen and gelatin. In other words, the reaction only changes the terminal chemical groups of collagen and gelatin molecules, rather than changing their molecular structure or conformation.^46,47 Therefore, the use of genipin was unlikely to induce denaturation of collagen I. Genipin, a naturally derived crosslinker obtained from the plant Gardenia jasminoides Ellis, has been used in traditional Chinese medicine for a long time.^48,49 It has been utilized as a biocompatible and cell-friendly crosslinker to replace synthetic crosslinkers with toxicity, such as formaldehyde and glutaraldehyde.⁵⁰ Therefore, it has been used as a routine and standardized crosslinker in drug delivery and post-treatment of biopolymer 3D printing in many studies.^6,51–59 Based on the CCK-8 test results in our study, MSCs proliferated over time, which confirmed that genipin was non-toxic to cells.

4.2. Effects of scaffold structure on cell proliferation

In this study, MSC proliferation was highest in sample 2 from day 1 to day 5. It had a smaller pore size than samples 3 and 4, which likely provided a larger surface area for cell attachment and improved the seeding efficiency and density of MSCs. Though sample 1 had the smallest pore size among all, its pores were even smaller than the rod diameter, rendering a low overall porosity. This limited porosity may have hindered the diffusion of solvents carrying water, nutrition, and metabolic waste, potentially affecting cell proliferation.⁶⁰

In short, our results were inconsistent with some existing studies regarding the effect of pore size on cell proliferation. For example, Sun et al.⁶¹ found that polycaprolactone scaffolds with a medium pore size (350 μm) promoted the proliferation of MSCs more than those with smaller (150 μm) and larger (750 μm) pores. Diao et al.⁶² found that tricalcium phosphate scaffolds with a pore size of 300 μm promoted MSC proliferation more than those with 100 and 500 μm-size pores. In contrast, in our study, the pore size at 144 μm exhibited the highest level of cell proliferation. We believe this difference was attributable to the type of scaffold biomaterials. Compared to rigid thermoplastic biopolymers and bioceramics, the hydrogel scaffolds used in this study were viscoelastic and hydrophilic, and therefore, more favorable for cell adhesion and migration.^63–65 Additionally, the collagen I composition had the surface RGD groups that guided and promoted cell adhesion, and it was also superior in nutrient adsorption during cell culture. Presumably due to these reasons, the use of the collagen I-based hydrogel could “switch” the optimum pore size to a smaller dimension.

In recent years, gelatin has been routinely used as a key component for producing hydrogel scaffolds, as it is a naturally derived protein with a surface RGD group like collagen.^{14,26,29,66–70} Though gelatin-based scaffolds have shown significantly improved outcomes in cell adhesion and proliferation compared to metal and thermoplastic biopolymers,^14,71 they remain inferior to scaffolds containing collagen I for bone tissue engineering. Compared to gelatin, collagen I is a native component of bone and has been shown to have a greater osteogenic-promoting effect.^5,72,73 Therefore, we achieved a high printing resolution with a hydrogel ink containing 8% w/v collagen I, which improved the biochemical and structural properties of the scaffolds and could synergistically promote osteogenic differentiation of MSCs. Therefore, our work steps further than recent studies utilizing gelatin-based scaffolds.

4.3. Possible mechanism related to the effect of scaffold structure on osteogenic differentiation

To our best knowledge, studies on the effects of structure on osteogenic differentiation are generally lacking in mesh scaffolds. In the few studies applying scaffolds with relatively big pore sizes (300–700 μm), the level of osteogenic differentiation increased monotonically in smaller pores within this range.⁷⁴ However, it is still necessary to produce scaffolds with much higher resolution to further explore their potential to promote osteogenic differentiation and investigate the possible underlying mechanisms.

In this study, the level of osteogenic differentiation was higher in sample 3 than in sample 4, indicating the advantage of using collagen I over gelatin. Among the collagen I-based scaffolds, sample 2 led to the highest level of osteogenic differentiation compared to samples 3 and 4. According to our WB test results, the expression of NCAD was the highest in sample 2, indicating the highest level of cell condensation and cell–cell communication, which could promote osteogenic differentiation, as reported in existing studies.^75–77 Additionally, the expression of NCAD was also the highest in sample 2 and maybe another contributor to promoting osteogenic differentiation.⁷⁸ Since MSC proliferation was highest in sample 2, it presumably led to a higher cell density, which promoted cell condensation and the formation of a hypoxic condition. These factors may have contributed to the upregulated expression of NCAD and HIF-1α. In addition, the expression of β-catenin was also highest in sample 2, and the upregulation of this factor has been shown to promote osteogenic differentiation in multiple studies.⁷⁹ We believe these findings could deepen our understanding of how scaffold dimensions and structures influence osteogenic differentiation and shed light on the possible mechanisms involved.

4.4. Broader impact of our study in bone tissue engineering

We believe the findings of this study offer several important contributions to the field of bone tissue engineering. First, our results provide insight into resolving the long-standing trade-off between cell proliferation and osteogenic differentiation in scaffold design.^62,68,74,80 Conventionally, smaller pores (100–150 μm) are known to promote osteogenic differentiation, but they often hinder cell proliferation. In bone tissue engineering, both factors are important: while higher proliferation increases the total production of ECM components, differentiation determines the proportion of bone-specific ECM components.^81,82 More importantly, our results indicated that higher cell proliferation may actually promote osteogenic differentiation by upregulating the expression of NCAD and HIF-1α. This highlights the importance of addressing both proliferation and differentiation in tandem rather than in isolation. In this study, such a trade-off was successfully solved in sample 2, which exhibited the highest levels of both proliferation and osteogenic differentiation, indicating that an optimal pore size can support both processes simultaneously.

Second, our work addresses a key challenge in scaffold development: the difficulty of optimizing both the compositional and structural properties for bone regeneration. Although collagen I is widely regarded as one of the most favorable biomaterials for bone tissue engineering,⁸³ previous studies have either failed to fabricate high-resolution collagen I-based mesh scaffolds or have achieved high-resolution scaffolds using alternative materials such as bioplastics, metals, or ceramics.^6,84,85 In our study, we addressed the printability limitations of collagen I-based hydrogels by introducing a Schiff-base interaction, which enhanced ink rheology and enabled the successful printing of high-resolution scaffolds. This allowed us to simultaneously optimize both composition and structure, enhancing the scaffold’s performance in supporting bone regeneration.

Lastly, much of the recent progress in bone tissue engineering has focused on biochemical enhancements, such as incorporating novel growth factors to stimulate osteogenesis.^86–89 In contrast, studies focusing on structural optimization have been relatively limited, presumably due to material processing constraints and insufficient scaffold resolution to promote osteogenic differentiation.¹⁴ While growth factors can be effective, their use often introduces added cost, complexity in formulation and release control, and potential safety concerns, all of which pose challenges for clinical translation.⁹⁰ Our results demonstrate that by achieving a high printing resolution with a collagen I-based scaffold, it is possible to promote both cell proliferation and osteogenic differentiation without relying on external biochemical agents. This presents a more straightforward and potentially more clinically translatable approach. We believe this study will inspire further exploration into the role of scaffold’s structural characteristics in promoting bone tissue regeneration.

5. Conclusion

In this study, we successfully addressed the limited printability of collagen I-based hydrogels by incorporating a Shiff-base interaction, enabling the fabrication of high-resolution scaffolds with rod diameters and pore sizes below 200 μm. This improved printing resolution was primarily attributed to an enhanced linear rheological characteristic of the hydrogel ink, rather than an increase in viscosity. We also identified an optimal combination of pore size and rod diameter that simultaneously promoted MSC proliferation and osteogenic differentiation. Mechanistic insights revealed that these effects were associated with the upregulation of NCAD, HIF-1α, and β-catenin expression. To conclude, we successfully designed and fabricated scaffolds with both optimized compositional and structural characteristics to support the proliferation and osteogenic differentiation of MSCs. We believe this study broadens our understanding of scaffold design and optimization, offering valuable insights for future applications in bone tissue engineering.