Recreating 3D limbal architectures by two-photon polymerization for cornea regeneration

Limbal Epithelial Crypts (LECs) are stem cell niches located in the limbus, the area between the cornea and the conjunctiva, and are responsible for cornea epithelium regeneration. Their anatomical structures were identified and described by Dua et al (2005). Stem cells migrate from this area toward the cornea where they differentiate into corneal epithelial cells. The lack of the LECs determines a pathology called Limbal Stem Cell Deficiency (LSCD) that can cause conjunctivalization, neovascularisation, corneal opacification, eventually resulting in vision loss. Currently, the main carriers to support limbal stem cell proliferation and differentiation used in clinics do not include enclosed structures able to act as cell reservoirs. In addition, the driving forces that determine the migration and the differentiation of limbal stem cells are not clarified yet. The hypotheses of this thesis were that the stiffness and the geometry of the scaffold influence the limbal stem cell differentiation process. Understanding these key factors could help to inform the design of scaffolds for future cornea regeneration strategies. The effect of the stiffness on cell differentiation is a well-known mechanism, but is less investigated in the limbus-corneal tissue. In this work, it was studied by producing macro-scaffolds (d=8 mm) obtained by UV-crosslinking. The stiffness was varied by increasing the concentration of Gelatin Methacrylate (GelMA), a biocompatible, photocrosslinkable material. Compression tests, rheology, AFM, and swelling analysis were performed on the developed hydrogels and compared to the ex vivo human cornea values. The gels with a concentration between 5% and 15% (w/v) exhibited mechanical properties in the same order of magnitude obtained for the cornea. The results indicated that the increased stiffness did not have a significant impact either in the expression of cytokeratin (CK3/CK14, stem cell marker, and CK14, differentiation marker), or in the gene expression (KRT3 and KRT19). However, the secondary key factors explored, differentiation media and oxygen concentration, supported cell differentiation. Inkjet printing was investigated as an additive manufacturing technique to produce 3D architectures. Although with this technique it was possible to develop a 3D structure with a stiffness gradient, its resolution was not sufficiently high to obtain a structure with dimensions comparable to the in vivo limbal stem cell niche. The second hypothesis was verified by developing a biocompatible scaffold mimicking the structure of the limbal stem cell niche and by evaluating the impact of its architecture on stem cell differentiation. The crypt geometry was modelled as U-shaped scaffolds with a diameter narrowing from 200 µm to 20 µm and was micro-fabricated from GelMA/PEGDA-based hydrogels using a Two-Photon Polymerization system (2PP). However, it was proven that the use of riboflavin as photoinitiator was inefficient at 780 nm, the wavelength used in the 2PP system. For this reason, P2CK was used as photoinitator to obtain stable hydrogels. The 2PP system allowed the precise recreation of the exact dimensions of the native crypts. Swelling, susceptibility to enzymatic degradation and stiffness were all evaluated. The biocompatibility of the printed scaffolds was assessed using immortalized human corneal epithelial cell proliferation up to 14 days. The ability of limbal stem cells to repopulate the crypts was demonstrated via two strategies. In the first strategy, human limbal stem cells were seeded directly inside the niche whilst in the second strategy, primary human limbal explants were placed adjacent to the printed structures and cells migrated towards the scaffold. Cell distribution and differentiation along the z-axis were evaluated using confocal microscopy. Cytokeratin 14 (CK14) with p63 and Cytokeratin CK3/12 (CK3/12) were used as limbal stem cell and differentiated corneal epithelial cell markers, respectively. Limbal epithelial stem cells were cultured in two conditions: xeno-free media, and with primary cells in serum containing media on a feeder layer. Both conditions showed the zonation of markers along the z-axis, which was not observed on flat scaffolds, demonstrating that the geometry alone influences cell phenotype. This suggests that the enclosed geometry results in the generation of a microenvironment inside the niche that influences cell phenotype. The presence of soluble factors, generated by cellular secretions, a specific oxygen concentration, and a more ‘stressful’ biomechanical milieu for the cells are some hypotheses that need further investigation and will be the basis of future work. In conclusion, the hypotheses of this thesis were partially confirmed. The variation in gel stiffness did not allow for the control of the hLESC differentiation process. However, the results demonstrated the influence of the geometry on stem cell differentiation without the use of signaling molecules. Further studies are necessary to have a description of the detailed spatial variability of the scaffold’s characteristics. Overall, the 2PP approach is flexible and could be applied to the generation of stem cell niches of other tissues, and could represent a significant advance in regenerative medicine