PCL Scaffold Combined with Rat Tail Collagen Type I to Reduce Keratocyte Differentiation and Prevent Corneal Stroma Fibrosis after Injury.

The cornea is one of the major refractive eye components with signicant functions, and its transparency is essential for clear vision. With regard to corneal injury, the corneal epithelium has a strong self-healing ability, while the corneal stroma is not capable of total self-repair. Therefore, preventing brosis and reducing keratocyte differentiation after injury have always been a challenge. The severe shortage of donor corneas for transplantation and transplant rejection prompted the development of corneal tissue engineering. In this study, we fabricated a poly(ε-caprolactone) (PCL) microbrous scaffold and infused the scaffold with rat tail collagen type I to obtain a 3D composite material. The PCL/collagen scaffold was designed to fabricate an optimal construct that simulates the stromal structure with properties that are most similar to the native cornea. The PCL scaffold has good mechanical properties, and infusion with rat tail collagen type I improved its biocompatibility. The results demonstrate that 3D composite material could reduce keratocyte differentiation, help achieve regular collagen distribution, and promote corneal repair.


Introduction
The cornea is a dome-shaped transparent tissue that functions as an optical lens that provides indispensable features to the visual system, including optical transparency and light refraction [1,2]. Meanwhile, the cornea is exposed to a constantly changing external environment as the outermost layer of the eye and is thus most likely to sustain damage; there are more than 1.52 million new cases of corneal blindness reported for each year [3], of which only less than 5% of patients have opportunities for corneal transplantation due to severe shortage of donor cornea and the high cost of transplantation surgery [4,5]. In terms of corneal injury, the corneal epithelium has a strong self-healing ability, whereas the corneal stroma may incur permanent corneal damage, leading to corneal stromal brosis and loss of vision [6]. Corneal transplantation has been the dominant procedure for more than 50 years, which highly depends on the availability of donor cornea [7]. Keratoprosthesis such as decellularized animal and human amniotic membrane transplant is associated with a lifetime risk of rejection [8,9].
To meet the demand for corneal transplantation, diverse bioengineering methods have been attempted to fabricate corneal substitutes based on synthetic (e.g., per uoropolyether (PFPE) [10], poly(vinyl alcohol) (PVA) [11,12], poly(ε-caprolactone) (PCL) [13,14], poly(methyl methacrylate) (PMMA) [15], poly(ethylene glycol) (PEG) [16], poly(hydroxyethyl methacrylate) (PHEMA) [17]or natural (e.g., silk [18], collagen [19,20], amniotic membranes [21], gelatin [17]) materials [22,23]. Furthermore, a combination of two or more synthetic and natural materials has been developed using 3D printing [24], electrospinning [13,17], and hydrogels [25]. The corneal stroma is the most central tissue of the cornea, making up two-thirds of its structure and accounting for about 90% of the corneal mass and thickness of the cornea [13]. The regular arrangement of collagen in the corneal stroma is fundamental to the transparency of the cornea; when the cornea is injured, quiescent keratocytes can be activated and transformed into broblasts and myo broblasts, which can secrete a large number of collagen bers randomly to repair damage, rendering the repair of corneal stroma damage di cult [26,27]. Therefore, the construction of a material that simulates the orthogonally aligned brous structure of the corneal stroma is expected to make new progress in corneal bioengineering.
PCL is a synthetic polymer that is widely used in the preparation of scaffold materials mainly because it is viscoelastic, ductile, isomer-free, and low-cost. In addition, PCL shows a certain degree of biocompatibility, long-term biodegradability, and absorbability [28,29]. It has been reported that the topological structure of a PCL ber-reinforced GelMA hydrogel can maintain the keratocyte phenotype and induce cell differentiation [30]. Studies have also shown that PCL scaffolds produced by electrospinning can be used to culture breast cancer cells in vitro [31]. These properties make PCL a good option for biological tissue engineering.
The PCL polymer is dissolved, and the solution is charged at a high voltage. When the electricity overcomes the surface tension, the polymer solution is pulled onto the target plate, and the solvent is evaporated to collect intersecting nano bers [32]. The structure mimics the stromal cell ber assemblage in vitro, and the cultured cells can take on a shape similar to that in vivo [33,34]. However, owing to the limitations of electrospinning in precisely controlling the network structure, the cells inoculated on the surface of the bers cannot migrate to the internal structure [35]. With the development of science and technology, near-eld electrospinning (NFES) was designed to overcome the intrinsic instability of traditional electrospinning processes and promote the controllable deposition of nano bers under the action of a reduced electric eld [36]. NFES or direct writing [37] can realize the hierarchical assembly and precise control of the network structure of sub-microscale bers [38]. Advances in technology have raised the prospect of making an alternative that more closely resembles a native cornea.
The corneal stroma is principally composed of aligned collagen type I brils principally [26]. Collagen, as a natural biopolymer, is widely used in biomaterials in the medical eld owing to its superior biocompatibility, biodegradability, and biological activity [39][40][41]. However, its poor mechanical properties limit its application in tissue engineering. Therefore, collagen is often used in conjunction with other materials to obtain mechanical support. Consequently, composite corneal grafts that possess mechanical properties and biocompatibility concurrently emerged [39,42,43]. The aforementioned PCL is an optimal choice. The PCL-collagen composite structure is applied not only in corneal tissue bioengineering but also in cartilage tissue engineering [41,44], heart valve tissue engineering [40], Achilles t5tendon repair [43], long-segment tracheal replacement [42], bone repair [45], and wound care [46]. The construct can simulate the complex corneal stroma better and supply a 3D environment for cell growth, providing a mechanical basis for cells to adhere, migrate, and proliferate, while collagen affords a suitable growth environment.
A synthetic substance is combined with a natural substance in order to complement the de ciencies of PCL and collagen and make full use of their advantages at the same time. In this study, we fabricated a multilaminate orthogonally aligned PCL scaffold precisely by NFES, which was perfused by collagen to form a 3D corneal graft. We selected the optimal composite scaffold by measuring the mechanical strength, light transmittance, water content, swelling ratio, biocompatibility, and topological structures. In vivo tests were conducted using a rat corneal stroma defect model to examine the brosis and regeneration of the corneal stroma after injury. Ultimately, we demonstrated that the designed composite scaffold has excellent mechanical and biological properties, which can reduce keratocyte differentiation and prevent corneal stroma brosis after injury.

Material And Methods
The materials and methods used in the study are detailed in the Supplementary materials and methods.

Results And Discussion
3.1 Fabrication of PCL/collagen scaffold. The PCL polymer was successfully used to fabricate the multilaminate orthogonally aligned scaffold by direct writing. Figure 1A and B show the actual and schematic representations of the direct writing equipment and the Taylor cone. In the process of direct writing, the formation speed of the jet and the movement velocity of the collector played an important role in controlling the morphology of the deposited bers. In our previously study [30], we found that a stable PCL jet was obtained under the following conditions: 3.6-kV electrostatic voltage, 20-kPa gas pressure, 3mm distance between the syringe needle and the collector plate, and a velocity of 60 mm/s. The grid scaffold was fabricated using a ber spacing of 200 μm and stacked with 10 layers. Many studies have reported the printing of PCL scaffolds by electrospinning technology; however, owing to the poor control of the prepared bers, only few studies fabricated orthogonally aligned scaffolds that successfully resembled the corneal stroma. They simply used PCL scaffolds as a mechanical support for corneal substitutes without a particular structure to simulate the corneal stroma [39,47,48]. It is di cult to control the pattern of PCL bers precisely by electrospinning; more importantly, the electrospun bers have low transmittance because of their high density, which makes them unsuitable as corneal substitutes. The NFES technique is a new electrospinning technology developed based on traditional methods to improve the control of bers [49,50]. PCL has a low melting temperature and stability, making it the most commonly used material in direct writing. However, PCL is less biocompatible, which limits its application in corneal bioengineering. Thus, we improved the biocompatibility of the PCL scaffold by infusing rat tail collagen type I.
The sterile rat tail collagen type I stock solution was neutralized with (NaOH 0.1M) to obtain a usable collagen solution at various concentrations ( Figure 1C). The concentration of 3 mg/ml was too low to gelatinize, while that of 5 mg/ml and 8 mg/ml was adequate for gelatinization ( Figure 1D). The PCL scaffold was fabricated by infusing collagen at concentrations of 3, 5, and 8 mg/ml mg ( Figure 1E). We cut the rectangular scaffold with a 3.25-mm-diameter trephine to obtain a corneal graft ( Figure 1F). Finally, we prepared the PCL scaffold and PCL/collagen scaffold for corneal transplantation ( Figure   1G). During the preparation process, collagen perfusion at a concentration of 3 mg/ml was unable to completely facilitate gelatinization, while collagen perfusion with a concentration of 8 mg/ml did not completely ll the interspace of the PCL scaffold due to its high concentration and poor uidity, resulting in uneven PCL/collagen scaffolds. For these aforementioned reasons, collagen at a concentration of 5 mg/ml was selected for perfusion. Orthogonally aligned PCL scaffolds infused with rat tail collagen type I were fabricated to simulate the native corneal stromal lamellae, which were rst used in corneal tissue bioengineering.
3.2 Macroscopic and microscopic morphology characteristics of the PCL/collagen scaffold. The design of the corneal stroma structure is pivotal to repair defects in the anterior corneal lamina due to its special structure. In this study, we constructed a PCL/collagen scaffold to simulate the complex orthogonal lamellar structure of the corneal stroma. It consists of a synthetic PCL scaffold and natural rat tail collagen, which complement each other mechanically and biologically. The morphology of the PCL bers and the deposition accuracy were observed by SEM at various magni cations ( Figure 2A). The PCL bers showed high deposition accuracy, with an average diameter of approximately 5μm. Horizontal and vertical printing constitutes one layer, stacking 10 layers of printing, making it 100 μm thick. Rat tail collagen has a loose and porous microstructure under SEM, which can provide a suitable growth environment for cells ( Figure 2B). The PCL scaffold was lled with collagen homogeneously without affecting the lamellar structure of the PCL scaffold, as observed by SEM ( Figure 2C). The collagen bers of the corneal stroma have a special orthogonal lamellar structure; however, the collagen secreted after injury is disorganized without its original structure, resulting in the di culty of completely repairing the corneal stroma. Some studies have also fabricated aligned bers, but they are only partially aligned [48,51]. The purpose of this grid-like PCL/collagen scaffold is to simulate the orthogonal lamellar structure of the native cornea to guide the regular arrangement of collagen bers after injury and repair the corneal stromal injury.
3.3 Physical characterization of the PCL/collagen scaffold. Transparency is essential for corneal transplant function and is associated with collagen tissue concentration, which contributes to transparency through its water-retaining proteoglycan [52]. The light transmittance of the cornea, collagen, and scaffolds were determined using a spectrophotometer at wavelengths ranging from 400 to 800 nm ( Figure 2D). Collagen at a concentration of 5 mg/ml and cornea had similar absorption curves in the visible light range, where the absorbance increased with increasing wavelength. The transparency of the cornea was approximately 87% [53], while that of collagen (64.74±8.11%) was lower than that of the cornea (87.77±6.19%) at 800 nm. The PCL scaffold with a ber spacing of 200 μm had a higher light transmittance (90.67±1.08%) than that of the cornea, but its light transmittance decreased upon collagen infusion. The wet PCL/collagen scaffold has a higher transmittance than the dry one owing to the hydroscopicity of collagen. As the water content of the graft increased, the transmittance increased correspondingly.
As mentioned previously, the cornea is a transparent water-containing tissue with the ability to moisturize; thus, the water adsorption and swelling properties of corneal substitutes play an important role in maintaining the structural stability of the corneal graft, which facilitates cell migration, adhesion, and proliferation, and delivers nutrients to the corneal cells. To further evaluate the physical properties of the scaffolds, the water content and swelling ratio were determined to identify the hydrophilic properties of the materials. In our study, the water content of the PCL/collagen scaffold with a collagen concentration of 5 mg/ml (85.26±0.94%) was not signi cantly different from that of the cornea (84.54±0.58%) ( Figure   2G). Furthermore, the swelling ratios of the PCL/collagen scaffold at various concentrations and the cornea were measured under the same conditions ( Figure 2H). No signi cant difference was observed in the swelling ratios of the cornea (32.66±2.29%) and PCL/collagen scaffolds with collagen at concentrations of 3 mg/ml (35.08±2.35%) and 5 mg/ml (31.12±0.61%). The water-uptake capability of the PCL/collagen scaffold with a collagen concentration of 5 mg/ml is similar to that of the native cornea, rendering it capable of meeting the requirement for use as a corneal substitute.
In addition to the transmittance and hydrophilicity of the PCL/collagen scaffold, the mechanical properties of the scaffold also play an important role in surgical suturing. Corneal substitutes require mechanical strength similar to that of native corneas or even higher in order to withstand the tension of the suture during surgery and the intraocular pressure after surgery. The tensile strain-stress curves and the maximum tensile strengths of the PCL scaffold, PCL/collagen scaffold, and rat cornea were measured. As shown in Figure 2E, the PCL scaffold had a higher elongation at break, which indicated that the ductility of PCL was higher than that of the native cornea. The elongation at break of the PCL/collagen scaffold and the native cornea was similar. The PCL scaffold and PCL/collagen scaffold attained the tensile strength requirements of the native cornea (2.06±0.16 MPa), with the maximum levels at 2.70±0.32 and 3.46±0.34 MPa, respectively ( Figure 2F). The results showed that the PCL/collagen scaffold was able to exhibit adequate mechanical strength to withstand both the surgical suture and intraocular pressure, which was mainly due to the mechanical properties of PCL itself, and this was enhanced by collagen infusion.
The mechanical strength, light transmittance, swelling ratio, and water content should be considered as design criteria in the development of corneal substitutes, as these characteristics are crucial to the cornea. Thus, we measured the characteristics of PCL scaffolds infused with collagen at concentrations of 3 mg/ml, 5 mg/ml, and 8 mg/ml to identify the optimal design for PCL/collagen scaffolds. As shown in Figure 2D, the light transmittance of the PCL scaffold was reduced by the infusion of collagen, but it could still conform to the requirement of the native cornea. From the results shown in Figure 2E and F, the maximum tensile strength of the PCL scaffolds was improved by collagen perfusion and was higher than that of the native corneas, meeting the requirement for suturing during surgery. Figure 2G and H show the hydrophilicity of the scaffolds, and we found that the water content and swelling ratio were inversely proportional to the collagen concentration. A higher concentration implies less space, limiting the storage and absorption of water.
3.4 Cell proliferation and adhesion properties of the PCL/collagen scaffold in vitro. Corneal substitutes require good mechanical properties and biocompatibility with no cytotoxicity. They should also permit cells of the injured tissue to adhere and migrate into the scaffold for injury repair and integration. Thus, cell adhesion and proliferation assays were performed on the scaffolds. In this study, we obtained limbal stromal stem cells (LSSCs) from the rat limbus, which can proliferate rapidly in vitro, and a Cell Counting Kit-8 (CCK-8) kit was used to quantitatively estimate the adhesion and proliferation of LSSCs ( Figure 3A). We selected the Matrigel commonly used in cell culture as the control group, and rat tail collagen with concentrations of 3, 5, and 8 mg/ml were used as the experimental groups. In CCK-8 experiments, the proliferation rates of LSSCs on the four plates were close to each other at 1 day of culturing, and there were no signi cant differences. The proliferation rates increased with incubation time, with the exception of collagen at a concentration of 8 mg/ml (Supplementary Figure S1). The proliferation rate in collagen with a concentration of 3 mg/ml was the closest to that in Matrigel regardless of culture time, followed by a concentration of 5 mg/ml. The proliferation rates in collagen at a concentration of 5 mg/ml showed no signi cant differences compared to that in Matrigel at 1, 3, and 5 days of culturing. It can be concluded that natural rat tail collagen is non-cytotoxic and can provide a good environment for LSSC proliferation; thus, the concentrations of 3 and 5 mg/ml of collagen were optimal.
The proliferation rates of LSSCs cultured in the PCL scaffold, collagen, and the PCL/collagen scaffold were also estimated ( Figure 3B). We observed that the proliferation rates of LSSCs on the three plates showed no signi cant differences after culturing for 1 day. The proliferation rates increased with the prolongation of culture time. The proliferation rate of LSSCs cultured in PCL scaffold was lower than that of the others, and the longer the culturing time, the more signi cant the differences were (Supplementary Figure S2). We speculated that the special structure of the PCL scaffold restricted the rapid growth of cells to some extent; it took longer for the cells to adhere and migrate to the regular PCL bers and proliferate along the bers. According to the differentiation of LSSCs and secretion of collagen after injury, the inhibition of cell proliferation may not be a disadvantage for injury repair.
The PCL/collagen scaffold provides more mechanical support, and the orthogonally arranged lamellar structure mimics the native corneal stroma and aims to guide the regular distribution of cells involved in repair and collagen secreted by broblasts and myo broblasts after injury to reduce brosis.
Immuno uorescence staining was used to observe the distribution of LSSCs on the PCL/collagen scaffold and on rat tail collagen without the PCL scaffold. LSSCs grew freely in collagen without speci c rules and with high cell density ( Figure 3C), while the cells grew along the two directions of PCL bers and were distributed on different layers on the PCL/collagen scaffold ( Figure 3D, E). In addition, the gap between the PCL bers provided ample space for cell growth.
In the scaffolds we fabricated, the concentration of collagen was found to affect the proliferation rates of LSSCs. At the same time, we con rmed that 5 mg/ml was the optimal concentration for the gelatinization of collagen. In addition, the PCL scaffolds can provide a topological structure for cells to adhere and migrate, which conventional tissue-engineered corneas cannot. Moreover, PCL/collagen scaffolds have microscale ber spacing rather than nanoscale spacing of electrospun brous scaffolds, which allows the 3D culture of LSSCs [54]. Compared to a conventional structural-free bioengineered cornea, it can be inferred from in vitro experiments that after corneal injury, the activated LSSCs can adhere and migrate along the PCL bers, and secrete extracellular matrix regularly to repair the corneal defect, so as to reconstruct the normal structure of corneal stroma. The pore of the scaffolds provides space for the growth and differentiation of LSSCs, which also preserves the function of cells to a certain extent and provides the basis for the repair and maintenance of corneal functions, while the traditional corneal replacement lls the corneal defect without retaining the cell function.
3.5 Histological analysis of rat corneal injury treated with PCL/collagen scaffolds in vivo. In our previous studies, we con rmed the mechanical properties and in vitro biocompatibility of PCL scaffolds and PCL/collagen scaffolds. Herein, we assessed the use of a PCL/collagen scaffold speci cally designed for corneal stroma wounds in a corneal injury model in Sprague-Dawley rats for the rst time. This was done by creating 60% deep corneal defects in SD rats, which was veri ed by H&E staining (Supplementary Figure S3b). The PCL scaffold and PCL/collagen scaffold were transplanted into wounded corneas through the keratoplasty, and the scaffolds were successfully sutured on corneas. Supplementary Figure S3a and c show the actual and schematic representation of the modeling process.
From the OCT images shown in Figure 4A, the red arrows indicate the positions of the transplanted scaffolds, which can be barely observed because of their thinness. We observed that the corneas exhibited severe edema, and the thickness increased signi cantly 1 week postoperatively. However, the corneal edema subsided with time, and the thickness of all corneas decreased within 4 weeks. The surface of the cornea became smooth 3 weeks postoperatively, except in the group with corneal damage, and the corneal interior of the two scaffold groups was denser than that in the group with corneal damage due to the regeneration of the corneal stroma. In the experimental group with PCL/collagen scaffold grafts, the corneal morphology gradually became normal within 4 weeks. However, at the end of our observation period, the surface of the cornea in the group with corneal damage was still smooth and the thickness was not uniform.
H&E staining was used to observe the secretion of collagen after corneal injury with different treatments ( Figure 4B). In H&E staining, the collagen bers appeared pale pink [55]. At the early repair stage, collagen secretion was abundant and irregular, accompanied by severe corneal edema and signi cantly increased corneal thickness. In the middle stage of repair, corneal edema subsided gradually, but the collagen bers secreted in the repair process were disorganized and without their original morphology and structure, causing corneal brosis. Four weeks after repair, with the degradation of rat tail collagen in the scaffold, the remaining collagen bers were arranged loosely and regularly; thus, the injured cornea treated with PCL/collagen scaffold graft transplantation showed a collagen arrangement most similar to that of the native cornea. Meanwhile, the untreated cornea showed a compact and irregular collagen arrangement. Histological H&E staining revealed that the PCL/collagen scaffold induced the regeneration of the stroma and regularly secreted collagen better than the PCL scaffold.
Collagen distribution and keratocyte differentiation were observed by immuno uorescence staining ( Figure 4C). Collagen VI is a unique feature of the extracellular matrix of corneal stromal tissue and is expressed in small amounts in normal corneal tissue (Supplementary Figure S4), secreted by broblasts after corneal injury [56]. Collagen is secreted in large quantities and arranged irregularly during the early stages of repair. With the passing of time and the subsidence of corneal edema, the collagen content gradually decreased, and its arranged structure also recovered. This was consistent with the results of the H&E staining. Vimentin is expressed speci cally in broblasts and is rarely expressed in normal corneas (Supplementary Figure S4) [57,58]. Thus, we evaluated keratocyte differentiation by observing the expression levels and distribution of vimentin. A large number of broblasts differentiated from keratocytes were observed in injured corneas throughout the whole stromal layer, especially in the early stages of repair. Four weeks after the operation, the broblasts and collagen bers observed in the corneas treated with scaffolds were signi cantly less than those in the untreated corneas; therefore, we speculated that the scaffold structure could inhibit keratocyte differentiation after injury. Meanwhile, the effect of inhibiting keratocyte differentiation in the PCL/collagen scaffold group was similar to that in the PCL scaffold group, possibly because the brous structure inside the collagen slowed down cell adhesion, migration, and transformation, leading to obstruction of cell growth.
Corneal thickness re ected the degree of corneal repair on the other side (Supplementary Figure S5). The corneal edema subsided, and the thickness of all corneas decreased 1 week after surgery. In the next 3 weeks, the thickness of all corneas decreased continually but was still slightly larger than that of the normal rats. Four weeks after repair, there was no signi cant difference between the thickness of injured corneas treated with PCL/collagen scaffolds and that of the untreated corneas.
Based on the previously mentioned results, the injured corneas treated with PCL/collagen scaffolds achieved ideal outcomes in the restoration of the corneal stromal structure, inhibiting keratocyte differentiation and corneal thickness increase, which is di cult to attain with conventional corneal substitution. Interestingly, we did not add biological factors to the PCL/collagen scaffold, indicating that the repair of the injured cornea depended on the topology of the scaffold. This feature not only helps prevent the immune rejection of the recipient but also makes the production, storage, and transportation of PCL/collagen scaffolds more conducive.
3.6 Related proteins and gene expression assessment of rat corneal injury treated with PCL/collagen scaffolds in vivo. We observed the expression of collagen VI (secreted after injury) and vimentin (speci cally expressed in broblasts) in the corneas by immuno uorescence staining, and the quantity of collagen VI and vimentin was measured by Western blotting. As shown in Figure 5A and B, vimentin was highly expressed in all corneas in the rst week of repair, and there was no signi cant statistical difference between the experimental groups. At the end of the 4-week observation period, the expression of vimentin in the PCL/collagen scaffold group returned to the normal level, while that of the group with corneal damage was still signi cantly higher than the normal value. The keratocytes transformed into broblasts after injury; nevertheless, the PCL construct inhibited the transformation, and the PCL scaffold infused with collagen showed a superior inhibitory effect.
Collagen is secreted in large quantities to ll the damaged area after corneal injury. However, we found an interesting phenomenon whereby collagen expression was signi cantly higher in the group with damaged cornea than in the groups treated with scaffold during the rst week of repair ( Figure 5A, B), which is the critical period for collagen production to repair the defect. Therefore, it is speculated that the scaffold lls the corneal defect area and occupies most of the space, resulting in reduced collagen secretion. The regular arrangement of collagen bers is a crucial factor in maintaining corneal transparency, which makes it necessary to reduce collagen secretion and guide the regular arrangement of collagen after injury. Various factors have been reported to regulate their bioavailability and function during corneal wound healing, such as transforming growth factor beta-1 (TGF-β1), TGF-β2, platelet-derived growth factor, and mesenchymal stromal cells (MSCs) [59,60]. However, for the rst time, we found that nonbiological scaffolds can also reduce keratocyte differentiation and collagen secretion during repair.
To further verify the effect of scaffolds with special structures on the repair of corneal stromal injury, we used whole-genome sequencing (WGS) and quantitative polymerase chain reaction (qPCR) to characterize the regulation of corneal wound healing. First, we screened out genes that were signi cantly altered after corneal injury by sequencing the whole genome of normal and injured corneas ( Figure 5C, D). In the volcano plot, the x-axis represents the log2 converted multiplier value of difference, and the yaxis represents the -log10 converted signi cance value. Differentially expressed genes (DEGs) are shown in the volcano plot as dots. Red represents upregulated DEGs, blue represents downregulated DEGs, and gray represents non-DEGs. In the cluster heat map, the horizontal axis represents the log2 of the sample, and the vertical axis represents the gene. The higher the expression level, the closer the color is to red, and on the contrary, the closer the color is to blue. Matrix metalloproteinases (MMPs) are signi cantly upregulated in injured corneas, which are involved in wound healing processes, including brosis [61][62][63].
Studies have shown that the upregulation of MMPs in diseased corneas is positively associated with levels of corneal neovascularization and brosis [64], which makes it meaningful to reduce the degree of MMP upregulation in inhibiting brosis. The gene expression of the proliferating cell nuclear antigen (PCNA), which re ects cell proliferation and MMPs, was assessed by qPCR.
From the results shown in Figure 5E, the expression of PCNA was upregulated in all corneas, with the expression being higher in the group with corneal injury, and there was no signi cant difference between the groups treated with PCL and PCL/collagen scaffolds. This suggested that the quiescent keratocytes were activated and differentiated into broblasts after injury, but decreased with time. The groups treated with PCL and PCL/collagen scaffolds showed lower proliferation rates, indicating that the special structure of scaffolds might inhibit the rapid proliferation of cells to a certain extent, which was consistent with the results of previous experiments.
Overexpression of MMPs has been associated with brosis in various tissues [65]. As shown in Figure 5F and G, the expression of MMP9 and MMP13 was signi cantly upregulated in all corneas. The expression of MMP9 showed no signi cant difference among the three groups in the rst week, but the expression of MMP9 in the group treated with PCL/collagen scaffold was lower than that in the other groups at the end of 4 weeks. The expression of MMP13 was similar to that of MMP9. These results indicate that the PCL/collagen scaffold transplanted into injured corneas can minimize the upregulation of MMPs, which can also reduce the transformation of keratocytes into broblasts and prevent brosis to a certain degree.

Conclusions
In summary, we fabricated PCL/collagen scaffolds by direct writing for the rst time, showing the capability to reduce the transformation of keratocytes and prevent brosis by inhibiting the upregulation of PCNA and MMPs. These scaffolds display a lamellar structure that simulates the native cornea.
Excellent mechanical properties and biocompatibility, as well as regulation of the adhesion and migration of cells to the PCL scaffold bers, have been demonstrated. We also discovered that the structure of the corneal stromal layer can be restored to the normal level, to the highest possible extent, after treatment of the corneal defect with PCL/collagen scaffold graft transplantation. Moreover, the transformation of keratocytes and the secretion of collagen were also inhibited, which is crucial for preventing brosis. This process was fully con rmed by OCT scanning, H&E staining, immuno uorescence staining, Western blot analysis, and qPCR analysis. The reduction of keratocyte differentiation and collagen secretion during repair by PCL/collagen scaffolds without biological factors has been investigated in animal experiments.
Using the topological structure of materials to promote wound repair provides a new option for corneal bioengineering and facilitates convenience in the preparation, preservation, and transportation of corneal substitutes. In addition, the advantages of non-immunogenicity and excellent biocompatibility of the PCL/collagen scaffold are highlighted throughout this study, showing a promising application potential in the eld of corneal bioengineering.   μm. (C) Immuno uorescence staining images of different groups after operation at each week. Green represents collagen type VI, red represents vimentin, and blue represents nuclei.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.