A Bioengineering‐Regenerative Medicine Approach for Ocular Surface Reconstruction Using a Functionalized Native Cornea‐Derived Bio‐Scaffold

Limbal stem cell deficiency (LSCD) is characterized by the loss of limbal epithelial stem cells (LESCs) which compromises corneal transparency, leading to blindness. It cannot be treated with pharmacological or corneal transplantation interventions, instead a specialized stem cell (SC) therapy is needed to restore eye health and sight. Herein, a native cornea‐derived biomaterial, a by‐product of a laser refractive surgical procedure called small incision lenticule extraction is identified as a new cell delivery matrix. Culture conditions are optimized to facilitate LESC attachment, expansion and stratification, and their identity is immunophenotyped. Using electron microscopy, bio‐constructs display stratification, similar to the architectural and cellular organization of a native mammalian cornea with formation of a basement membrane and an orderly array of collagen fibrils. Neuronal growth and depleted CD45+/CD14+ leukocytes on lenticules are also shown, suggesting that in transplantation experiments, they will re‐innervate and not trigger a host‐mediated immune response. Finally, human lenticules are geometrically customized to successfully fit them over a LSCD murine cornea ex vivo, during which they maintain curvature. The authors are poised to conduced similar studies in live mice using these and other carriers currently used in the clinic to compare SC therapy outcomes.


Introduction
[3] Limbal stem cell deficiency (LSCD) is a debilitating disease of the ocular surface characterized by loss of limbal epithelial stem cells (LESCs), resulting in a pathological wound healing response known as "conjunctivalization" which compromises corneal transparency and leads to blindness. [4]LSCD cannot be treated with pharmacologicals or a corneal transplant; instead a specialized stem cell (SC) therapy is deployed to restore eye health and sight. [5][8][9] It is difficult to pin-point the reason(s) why failures arise, but disease severity, duration, etiology, genetics, co-existent ocular disease, type of therapy, and source of SCs and their mode of transplantation, likely influence outcomes.The other key variable to consider is the material used to support and transport SCs to the ocular surface.Such a matrix should entice cell attachment, expansion, and maintenance of stemness, while regenerating the cornea to restore its form and function.
Various biological and synthetic scaffolds have been used for SC transplantation, including human amniotic membrane (hAM), fibrin, human anterior lens capsule, chitosan, collagens, keratin, silicon hydrogel contact lens, and others. [10]Among them, hAM and fibrin have been used in human clinical trials.However, hAM is costly because of microbial screening and storage, it is difficult to access and has suboptimal optical and physical properties. [11,12]Fibrin is a blood product, is difficult to handle, maneuver, and suture over the recipient cornea and has been associated with parvovirus B19 infection and anaphylaxis in general surgery. [13,14]Furthermore, neither biomaterial is cornea-derived, thus their biocompatibility and integration capacity is questionable.Alternatively, synthetic carriers such as US Food and Drug Administration-approved siloxane-hydrogel contact lenses have been used to deliver LESCs with some success. [15]This therapeutic option can be xeno-free, autologous, and cost-effective.However, at some point during cell transfer, the device must be removed; despite the added protection it offers donor cells and the recipients' cornea during engraftment.
Small incision lenticule extraction (SMILE) is a laser refractive surgical procedure to treat myopia (short-sightedness), astigmatism (irregularly shaped cornea), [16] and keratoconus. [17,18]It uses a femtosecond laser to excise and extract an intrastromal layer of corneal tissue with a diameter of ≈7.0 mm and a thickness of 50-140 μm (Figure 1a).This surgical by-product is typically discarded as medical waste.Lenticules are generally extracted from otherwise normal corneas and are readily available given the procedure has already been performed on millions of individuals.Previously, they have been implanted for correcting hyperopia [19,20] and other refractive errors, [21] or used as a patch for treating corneal perforation, [22] with no adverse side effects reported, indicating they can be safely transplanted with low risk of rejection.The novelty that our study brings to the field includes the use of lenticules to carry LESCs for corneal epithelial rehabilitation.Therefore, our short-term goal was to determine whether SMILE lenticules can be adapted as a nurturing and transfer medium for healthy donor cells to diseased corneas.The main objective was to identify the culture conditions necessary for carrying corneolimbal epithelial cells (CLECs) on this biological substrate and determining its feasibility for transplantation.Herein, we optimized CLEC attachment, expansion, and stratification, and determined morphological, phenotypic, functional, and structural features of this bio-construct in 3D organ culture.We also determined how to fit and fix them on wounded murine corneas in mock transplantation experiments.With further refinements, this could be developed into a new solution for regenerating the ocular surface in LSCD.

Results and Discussion
In addition to disease heterogeneity due to the numerous causes and cellular sources that have been considered as treatment options for LSCD, for which there is still no clear cut gold standard, [23][24][25] the other important variable that likely dictates therapeutic outcome is the substrate used to carry SCs to dis-eased corneas.Systemic injection of SCs in organs that need replenishment is controversial as it is estimated that less than 5% reach their target. [26,27]In the realm of repairing injured corneas, systemic or subconjunctival administration of mesenchymal SCs can reach their intended destination, [28,29] although some researchers believe this is not the case. [30]Therefore, in the cornea, as with many other organs, scaffold-based cell delivery modalities can overcome these shortcomings.Fibrin and hAM are two efficacious biological carriers used to graft LESCs for rehabilitating the cornea and restoring vision in patients with LSCD. [31,32]here are certain features that render a biomaterial suitable as a substrate-carrier for SCs.Notably, it should provide an appropriate microenvironment, including a biological framework that encourages cell adhesion, expansion, re-organization, and restoration of tissue function.If the scaffold integrates into the host site and offers equivalent or better biological and bio-mechanical properties compared to the native tissue without eliciting an adverse reaction, then these are advantageous attributes that should be considered in developing tissue engineering and regenerative medicine-based strategies for the cornea and potentially other tissues.In the biomaterials space, numerous applications have been devised which have not yet reached the clinic but are relevant nonetheless, including 3D printing of limbal architecture with extracellular matrix proteins (ECM) proteins, bio-active hydrogels, and decellularized corneas. [10,25]hen devising cell-based therapies for patients with LSCD, it is important to acknowledge that the limbus is often severely damaged, and this is concomitant with SC depletion in the same zone.Hence, the ultimate treatment would be one that addresses both issues.[35] In clinical trials to treat LSCD, there have been no genuine attempts to recreate the limbus; most (if not all) options rely on repopulating the depleted SC pool.It is questionable whether these strategies will restore limbal architecture. [36]Therefore, our aim was to ensure SMILE lenticules supported CLECs including SCs, to facilitate re-epithelialization and healing in LSCD, and not an attempt to replace the limbus, although if limbal architecture is corrected, this would be deemed a significant advancement in the realm of ocular surface rehabilitation and vision restoration for this disease.

CLECs Cultured on SMILE-Derived Lenticules
Advances that have focused on the use of native corneal tissue as a scaffold-based cell delivery system [37] inspired us to pursue and further progress this area of research.To this end, our attention turned to the by-product of a refractive surgical procedure known as SMILE and the ultra-thin lenticules extracted from the corneal stroma of individuals that undergo this elective procedure.Given that at least 5 million SMILE surgeries have been conducted worldwide (https://eyewire.news/news/zeiss-marks-milestonewith-more-than-5-million-eyes-treated-with-smile?c4src = article:infinite-scroll) and the utility of lenticules in the treatment of a variety of ocular diseases, [19][20][21][22]38] there is growing interest in developing preservation protocols [39,40] and banking systems [41,42] for this valuable tissue resource.
Therefore, we hypothesized that lenticules derived from native human corneas after SMILE refractive surgery (Figure 1a) possess favorable physical and biochemical properties for propagating CLECs for corneal regeneration after transplantation.To this end, lenticules collected immediately post-surgery and stored at 4 °C in saline displayed remarkable transparency (Figure 1b, first panel), even after 7 days of cultivation with human corneal epithelial cells (hCECs) in defined keratinocyte serum-free complete media (dKSFM) (Figure 1c).Notably, cell density was low, and they displayed an enlarged and rounded morphology, as if detaching (Figure 1c, arrow and arrowhead, respectively), which suggested that the substratum was less than ideal for their attachment and expansion.
To determine whether this was due to the cell-type used, other strategies were trialed including single cell suspensions and tissue explants.hCECs (Figure 1d) and primary Confetti mouse CLEC (p-mCLECs) (Figure 1e) also unsuccessfully colonized the lenticule surface, although some pan-cytokeratin + , p63 + , and Confetti + epithelia were detected.Interestingly, hCECs were found on both the anterior (Figure 1d-i) and posterior (Figure 1d-iii) surface of the lenticule, while none were detected within the keratocyte-containing stromal layer (Figure 1d), suggesting that they do not burrow into the lenticule matrix.Precisely how they achieved this was not determined, although we speculate that hCECs migrated from the anterior to the posterior aspect or detached from the tissue culture plastic prior to adhering to the posterior side.Using Confetti p-mCLEC as a means of monitoring cell colonization in real-time, similar results were obtained (Figure 1e-i-iii).Finally, Confetti limbal tissue was procured and placed on the lenticule surface, rationalizing that SC maintenance signals will diffuse from explants to facilitate expansion and colonization. [7,43]Phase contrast (Figure S1, Supporting Information, upper-first panel) and fluorescence (Figure S1, Supporting Information, lower-first panel) microscopy identified an initial outgrowth after 5 days in culture.Upon removing the explant, an area of fluorescing Confetti epithelia remained attached to the lenticule surface (Figure S1-i, Supporting Information, upper-second panel), some of which exhibited the capacity for limited expansion (Figure S1-i, Supporting Information).Red fluorescent protein (RFP) and yellow fluorescent protein (YFP)expressing cells were confined to the upper surface of the lenticule (Figure S1-i, Supporting Information) and were not detected on the underside or indeed within the lenticular matrix (Figure S1-ii,iii, Supporting Information).However, epithelial coverage was not achieved under these conditions, suggesting they were less than ideal.

CLECs Cultured on Coated Lenticules
In realizing their translational potential for ocular surface reconstruction in LSCD, we dispensed mammalian corneolimbal epithelia on lenticules, and although our results were encouraging in that some cells initially adhered, expansion and colonization were not as forthcoming as anticipated (Figure 1).This was not unexpected given the laser-assisted dissection likely denatures key ECM and adhesion proteins within the lenticule matrix. [44]herefore, we tested whether coating lenticules with biologicals that either simulate or comprise the epithelial basement membrane, promoted cell adhesion, and expansion.Matrigel was used as it contains a host of proteins, including laminin and collagen type IV, which are found on the limbal basement membrane [45] and used as a 3D scaffold to model the limbal niche. [46]Fibrin is a blood component which has been adopted as a transplantation medium for LESCs in preclinical models and clinical trials for treating LSCD. [7,47]Upon testing Matrigel and fibrin as coating matrices, Confetti p-mCLEC adhered on the first day of seeding (Figure S2a,b, Supporting Information, first column).However, colonization of the lenticule surface was unsuccessful during the ensuing 7 days (Figure S2a,b, Supporting Information, second column).Finally, lenticules were incubated in FBS as it contains a myriad of growth factors and adhesion molecules, including fibronectin and vitronectin. [48]However, p-mCLEC expansion on 10% FBS-coated lenticules was also ineffective (Figure 2a,b).Instead, numerous irregular-shaped phalloidin + cells were identified within the lenticular matrix (Figure 2a).A magnified view of a randomly selected region disclosed that these cells were resident lenticular cells that displayed a stellate-spindle morphology, consistent with keratocytes.Only a few superficial larger cells were found attached which exhibited prominent nuclei and broader cytoplasmic actin distribution (Figure 2a-i, arrows), features consistent with epithelial cells, especially given their K14 expression (Figure 2b).To discern the effect of FBS on keratocytes, lenticules were cultivated in dKSFM media containing 10% FBS for up to 3 weeks (Figure 2c).Staining for phalloidin identified a meshwork of fibroblasts (Figure 2c-i), most of which were proliferating as denoted by Ki67 + cells (Figure 2c-ii).In contrast, lenticules that were not exposed to FBS harbored few if any healthy fibroblasts (Figure 2c-iii).Under these conditions, fibroblasts had the propensity to emigrate from the lenticule and colonized the tissue culture plate (Figure 2c-iv).Thus, to minimize keratocyte presence and proliferation, and maximize epithelial cell attachment and expansion, lenticules were cultivated in dKSFM with lower FBS content for up to 2 weeks.Under these conditions, few if any keratocytes were detected in media supplemented with either 1% or 5% FBS (Figure S2c, Supporting Information).
Next, to ensure any remaining keratocytes were quenched, lenticules were ɣ-irradiated.For this arm of the procedure, lenticules were initially incubated in dKSFM media containing 5% FBS for 2 weeks, ɣ-irradiated at 1000cGy, briefly washed in phosphate buffered saline (PBS), and further incubated in media without FBS for 2-3 days prior to seeding hCECs on lenticules (Figure 2d).Under these conditions, epithelia successfully attached and expanded, and there was no difference in their proliferation status (30.67% ± 2.082 vs 41.67% ± 10.016, p = 0.136; irradiated vs non-irradiated, respectively) or morphology between irradiated and non-irradiated lenticules, as visualized by Ki67 and phalloidin staining (Figure 2d).Notably, these results are contrary to what occurs on type I collagen and fibrin-coated plastic substratum, [49] and in 3D culture of the same cells with Matrigel and hAM extract. [46]Yet other researchers have reported that CECs fail to adhere to Matrigel-coated collagen shields, while they stratify on type IV collagen-coated surfaces. [50]In this situation, instead of promoting expansion of donor epithelial cells, we noted significant growth of resident lenticule keratocytes (Figure 2c).An additional concern was that keratocytes would differentiate into fibroblasts, and this might contribute to excessive matrix remodeling including loss of transparency and opacification (Figure 2c-iv).To counter this effect, we decreased the concentration of FBS to 1-5% (Figure S2c, Supporting Information) and ɣ-irradiated bio-constructs to prevent keratocyte proliferation (Figure 2d and Figure S2c, Supporting Information).
Admittedly, our protocol required a series of modification before epithelia flourished on the lenticule surface.Notably, fresh from SMILE surgery, lenticules harbors no epithelium and endothelium, and are devoid of a BM on either side.In some ways they represent tissue that has endured a severe injury.Thus, seeding corneolimbal epithelia on them is likely detrimental to their viability as they are now in direct contact with keratocytes and exposed to stromal proteins that they would normally be segregated from until they recapitulate their own BM. [51,52]

Cultivation and Characterization of CLECs on Irradiated Lenticules
Having determined the optimal conditions for suppressing keratocyte proliferation and/or rendering them non-viable within the Human and mouse corneolimbal cells cultured on coated SMILE lenticules.a) Lenticules were coated with 10% FBS, and a single cell suspension of p-mCLECs were seeded on them for 7 days (n = 3) after which they were stained for phalloidin (red) and Hoechst (cyan).The hatched white square is magnified in panel i (right).Scale bars, 700 and 50 μm.b) Lenticules were coated with 10% FBS, and a single cell suspension of p-mCLECs was seeded on them for 7 days (n = 3) after which they were stained for K14 (red) and Hoechst (blue).The hatched white square is magnified in panel i (right).Magnified orthogonal views through the cross-section indicated by the dotted yellow line.The dashed white lines in the orthogonal views demarcate the anterior and posterior surfaces of explant (i, right).Scale bars, 700 and 50 μm.c) Lenticules cultured in dKSFM containing 10% FBS for 3 weeks (n = 3) and stained for phalloidin (green)/Hoechst (blue) or Ki67 (violet)/Hoechst (blue) (i and ii, respectively).Specimens cultured in the same media without FBS were used as control (iii) (n = 3).Representative phase contrast image of fibroblasts on expanding on tissue culture plastic which emigrated from a lenticule during 3 weeks in the dKSFM media containing 10% FBS (iv).Scale bars, 50 and 100 μm.d) Lenticules conditioned in dKSFM containing 5% FBS for 2 weeks and ɣ-irradiated at 1000cGy, cultured in FBS-free dKSFM for 2-3 days, then seeded with hCECs (left panel; n = 3).Lenticules cultured in the same condition without ɣ-irradiation were used as controls (right panel; n = 3).Cells were stained for phalloidin (green), Ki67 (purple), and Hoechst (blue).Scale bars, 40 μm.lenticule matrix, we next sought to determine the effectiveness of this protocol on CLEC expansion (Figure 3a).When hCECs were seeded on lenticules (Figure 3b), they attached, expanded, and formed large circular colonies (Figure 3b-i), some of which displayed basal K14 expression in the center (Figure 3b-ii).hCECs rapidly expanded, forming multiple layers (Figure 3b; Movies S1 and S2, Supporting Information) with a mean lenticule coverage of 54.25% ± 26.75% after 10 days.Superficial cells were minimally differentiated, denoted by negative expression of K12 (Figure 3b-iii), while those in the basal layer displayed heightened K14 expression (Figure 3b-ii).Notably, no significant keratocyte expansion was detected within the lenticule matrix (Figure 3bii,iii and Movies S1 and S2, Supporting Information).
Similarly, p-mCLECs formed large circular holoclone-like colonies (Figure 3c-i), which have been deemed to harbor SCs. [53]owever, as expected for primary cells, their growth was significantly slower compared to hCECs (Figure 3b).The mean lenticule surface area covered by these colonies was 26.82% ± 22.01%.Again, multiple layers formed including cells that were less differentiated, as indicated by their K12 − /K14 + content (Figure 3cii,iii), suggesting progenitor-like cells were expanding on the construct.Given K14 has been deemed a marker of LESCs, [54][55][56][57] this is further evidence that SMILE lenticules support limbal cells including a population of precursors.Moreover, few if any keratocytes were detected within the lenticule matrix (Figure 3c-ii,iii and Movie S3, Supporting Information).Under these conditions, epithelial cell morphology was maintained, and these cells retained their proliferative capacity, as denoted by Ki67 immunoreactivity (Figure 3d).
After consolidating and optimizing our strategy, we ascertained that CLECs successfully expanded with the emergence of large circular holoclone-like colonies from hCECs and p-mCLECs, indicative of progenitor cell activity (Figure 3) and the propensity for self-renewal (Figure 3b-d, and Movies S1-S3, Supporting Information).Certainly, this data aligns with recent reports using embryonic stem cells (ESCs) and induced pluripotent stem cell (iPSC)-derived CECs expanded on lenticules. [58,59]

Stratification of Epithelia on SMILE Lenticules
Due to their rapid expansion and stacking potential (Figure 3b), hCECs were used to induce stratification via an air-lifting protocol (Figure 4a).Transmission electron microscopy (TEM) demonstrated that hCECs successfully laminated into three distinct types including basal, wing, and superficial squames (Figure 4bi).In addition, an electron dense basement membrane-like structure formed beneath basal epithelia (Figure 4b-iii), and collagen bundles within the lenticule matrix remained highly organized as orthogonally aligned fibrils (Figure 4b-iv).Scanning electron microscopy (SEM) identified extensive apical projections consistent with microvilli and the occasional desquamating superficial cell (Figure 4b-ii), equivalent to what transpires on a native mammalian cornea. [60]Laminating epithelia expressed tight, adherens, and gap junction proteins including N-and P-cadherin, occludin, and connexin-43 (Figure 4c), similar to the profile displayed in cells cultivated on human amniotic membrane [61] and comparable to those which populate a native cornea, [62,63] sug-gesting that barrier function was readily established in this 3D tissue culture system.

Immunogenicity of Lenticules and Cultivation of Trigeminal Ganglion (TG) Neurons
To minimize donor antigenicity and genetic material being carried to the recipient, de-cellularization and de-nucleation protocols have been applied to freshly sourced lenticules. [64,65][68][69] Certainly, immune cells have been detected in fresh lenticules compared to few or none in preserved specimens. [70]We confirmed this notion by identifying a population of CD45 + leukocytes within lenticules, further specifying them as CD14 + macrophages (Figure 5a, first-third columns).Importantly, once exposed to our conditioning protocol, there were few if any detectable CD45 + /CD14 + cells (Figure 5a, lower first-third panels), implying we have lowered their immunogenicity.
Key to integration and restoration of function is reinnervation of the donor bio-construct.Although it has been documented that nerve axons are detectable in fresh and cryo-preserved lenticules, [71] the refractive procedure severs them from their cell body, rendering them non-functional.The cornea is one of the most innervated tissues of the body. [72]It is thus inevitable that the femtosecond laser that cuts the corneal stroma in two parallel planes, severs numerous axons that course through it, thereby causing denervation in that region. [73]The level of denervation is dependent on the lenticule thickness, that is, the amount of refractive error to be corrected.If SC-laden lenticules are to survive and restore tissue function, then reinnervation from the recipient is necessary.Yam and colleagues presented proof-of-concept that this can be achieved ex vivo by grafting chick dorsal root ganglia onto decellularized lenticules. [74]e identified thick stomal nerve axons immediately post-SMILE surgery (Figure 5a, upper last column) including their deterioration with time in culture (Figure 5a, lower last column).This makes perfect sense given their dislocation from the nerve body in the TG.We also demonstrated that neurons isolated from mouse TG successfully adhered to the lenticule surface, and with time spawned a network of III-tubulin + neurites (Figure 5b; Movies S4 and S5, Supporting Information).Overall, neurite length (35.30 ± 31.32 vs 2.21 ± 2.54, p = 0.034) and density (8.52 ± 7.80% vs 0.413 ± 0.36%, p = 0.035) significantly increased at day 4 compared to day 0 (Figure 5c).
In considering these bio-constructs for grafting in preclinical models of LSCD, we prepared mini lenticules (without cells)  sloughing superficial cells from the bio-construct (ii), formation of an electron dense basement membrane (BM) (iii), and aligned collagen bundles within the lenticule matrix (iv) were some characteristic features imaged.Scale bars, 2 (i), 5 (ii), 100 (iii), and 1 μm (iv).c) Stratified hCECs on lenticules were stained for N-cadherin (red), P-cadherin (red), occludin (red), and connexin-43 (green) with phalloidin (green or red) and Hoechst (blue).An IgG isotype antibody was used as a reagent control for which no immunoreactivity developed (n = 3/group).Scale bars, 20 μm. of different geometries including 2 and 3 mm circular and 1.5/3 mm annular-shaped specimens by trephination (Figure 6a).As proof-of-concept, they were placed on corneas of euthanized mice that had undergone total mechanical debridement of the limbus and central cornea (Figure 6b) or total mechanical debridement of the limbus and peripheral cornea leaving the central region intact (Figure 6c).Optical coherence tomography (OCT) with real-time fundus view showed that each design was a snug fit over the recipient cornea, and importantly maintained curvature, indicating their feasibility for ocular surface reconstruction in LSCD.However, an obvious limitation is their thickness, especially when implanted on a mouse cornea.To generate thinner alternatives, excimer laser was used to generate specimens of ≈25 μm thickness (Figure 6d,e).Ablated lenticules were a better fit compared to their unattenuated counterparts and were thinner than the total thickness of native mouse cornea (Figure 6b,c).
Using the ablation modality, Damgaard et al. also demonstrated that the lenticule surface is smoother compared to those obtained after SMILE surgery. [44]Using this approach, it is possible to standardize lenticle size, shape, and thickness, especially if deemed an appropriate strategy to treat LSCD.Although we succeeded in reducing lenticule thickness by excimer ablation by ≈75-80%, using OCT they still appeared too thick (Figure 6d,e).Reducing their dimension further is technically challenging, however thickness is also likely to be influenced by edema which develops while the lenticule is stored in saline post-extraction or following conditioning during the cultivation phase.We suspect that once successfully integrated into the host cornea, this swelling will subside.
Certainly, our bio-constructs can be reshaped and resized to fit a mouse cornea (Figure 6).However, in considering this as a future therapeutic strategy for patients with LSCD, they may be too small to cover the surface area of the human cornea.Therefore, we envisage three transplantation options, each potentially requiring more than one constructs (Figure 7).Option 1 is the typical strategy whereby the cornea is completely covered by the grafted material.Option 2 is analogous to a technique called simple limbal epithelial transplantation, whereby donor limbal tissue is diced and attached to the recipient cornea in a random fashion. [79]Option 3 is a rather novel approach in that the graft is placed over the limbus, where LESC typically reside.This third modality is designed to encourage engraftment in a peripheral location and ensures the visual axis in not obstructed (Figure 6).For each option, the graft can be glued or sutured in place, then covered by a therapeutic contact lens to protect the donor material during the implantation phase.

Conclusion
In reaching our goal on the use of SMILE lenticules as a carrier for LESCs, our future aspiration will be to develop a better treatment for LSCD.Lenticules will be the foundation of the bioconstruct, harvested from a donor, and not the individual with LSCD.Hence, this portion of the graft would be deemed allogeneic.The second component of the bio-construct includes SCs, which can be sourced from the patient with LSCD if they present with unilateral disease, or from a cadaver or relative, rendering this component of the graft autologous and allogenic, respec-tively.Irrespective of the SC source, it may be necessary for recipients to receive some form of immunosuppressive therapy to minimize the risk of rejection.
Apart from using lenticules as a carrier for LESCs to treat LSCD, our constructs may have great utility as 3D tissue equivalents for drug screening and toxicity evaluation, wound-healing and cell-to-matrix interactions, and used as an advanced alternative to 2D monolayer cultures or expensive animal models.[82] In the future it will be important to compare this new biomaterial with those that are in current clinical use to determine whether clinical outcomes can be improved.

Experimental Section
Ethics Clearances: UNSW Human (HC15647) and Animal (20/59A) Research Ethics Committee approvals were obtained for tissue acquisition and experimentation.Protocols associated with UNSW approvals were also reviewed and approved by the U.S. Army Medical Research and Development Command, Office of Research Protection, Human Research Protection Office, and the Animal Care and Use Review Office.Corneal tissue was also obtained from lifeact mice as part of this institutional tissue sharing arrangement.
Human Donor SMILE Lenticule Acquisition and Preparation: Fresh lenticules (left and right-eye; n = 113) were obtained from consenting patients (both female and male; age range: 21-50 years) that elected to undergo SMILE laser refractive surgery at the Envision Eye Centre (Sydney, Australia).Lenticules were collected and kept in sterile saline at 4 °C until experimentation.They were conditioned to receive CLECs by submerging them in dKSFM (Thermo Fisher, Waltham, MA) comprising 10 ng mL −1 epidermal growth factor (R&D Systems, Minneapolis, MN), 100 ng mL −1 cholera toxin (List Biological Laboratories, Campbell, CA), and 100 U mL −1 penicillin-streptomycin (Thermo Fisher) for 2 weeks at 37 °C in 5% CO 2 , otherwise referred to as complete media.Lenticules were used either immediately after conditioning or following gamma (ɣ)-irradiation with 1000cGy for 20 min.Next, they were washed twice in PBS and further incubated in complete dKSFM media without FBS for 3 days prior to receiving a suspension of cells or tissue explants for subsequent cultivation.This strategy was employed to mitotically inactivate and/or deplete viable lenticule-associated keratocytes.
Functionalizing Lenticules by Coating with Biologicals: To entice cell adhesion and expansion, recombinant and native ECM proteins were coated on the lenticule surface.Fresh surgical specimens that were originally stored in sterile saline, were incubated with i) 10 μg mL −1 Matrigel (Thermo Fisher) for 2 days at 4 °C, ii) 10 μL of fibrin comprising fibrinogen and thrombin according to packaging instruction (TISSEEL; Baxter, Deerfield, IL) overnight at 4 °C, iii) 200 μL of 2.5 μg mL −1 recombinant human vitronectin (RayBiotech) overnight at 4 °C, or iv) 1-10% FBS (Thermo Fisher) for 3 h at 37 °C.Any excess coating solution was removed, specimens rinsed in PBS and limbal tissue explants or a suspension of CLECs (2-5 × 10 5 ) seeded on the lenticule surface.
Culturing Human and Murine Corneolimbal Epithelia: Cells were originally expanded using tissue explants from healthy C57BL/6, K14CreER T2 -Confetti and Lifeact mice.Four weeks-old mice were euthanized, their eyes enucleated, sterilized in 2% iodine (Sigma-Aldrich Corp., St. Louis, MO), and dissected corneas placed in culture for epithelial cell expansion. [47]riefly, the sclera, conjunctiva, and iris were removed, and the tissue was orientated epithelial side down in 6-well plates (Corning Costar, Sigma-Aldrich) for 10 min at room temperature to facilitate attachment.Next, 100 μL of complete media (without FBS) was added, ensuring the explant was not disturbed, then incubated for 1 h at 37 °C in 5% CO 2 .Additional media was introduced, and cultures maintained long-term at 37 °C in 5% CO 2 .Media was exchanged thrice weekly, explants were removed after  Potential SC-lenticule grafting options for patients with LSCD.We envisage that lenticules will be seeded with autologous or allogeneic LESCs.When grafts are deemed ready, patients will be scheduled into theatres to have a superficial kerectomy, that is, removal of the opaque conjunctival pannus from the cornea.After which three transplantation options could be employed.Option 1: Use several cell-loaded lenticules to cover the diseased cornea.Option 2: Use single or multiple cell-laden lenticule, cut into small wedges which are randomly distributed on the cornea.Option 3: Use single or multiple cell-laden lenticules, cut into small concentric annular-shaped strips which are place around the perimeter of the cornea.Created with BioRender.≈10 days, and emerging cells passaged once they reached 80% confluence by brief exposure to 1X trypsin-EDTA (Thermo Fisher).In some experiments, corneolimbal explants were placed directly on the lenticule surface and removed after 7-10 days.K14CreER T2 -confetti mouse corneas were used for their endogenous expression of multiple fluorescent proteins including cyan fluorescent protein, green fluorescent protein (GFP), YFP, or RFP in keratin (K)−14 + limbal cells.Likewise, lifeact murine CLECs were also applied to the lenticule surface due to their expression of enhanced-GFP in F-actin filaments, thereby allowing the authors to trace their fate, that is, morphology, organization, and distribution.SV40-transformed human corneal epithelial cells (hereinafter referred to as hCECs) were cultivated under the same condition.These cells have been authenticated and comprehensively characterized. [83]solating and Cultivating Murine TG: Murine TGs were isolated as previously described for dorsal root ganglion. [83]Briefly, TGs from healthy 6-8 weeks-old C57BL/6 mice were dissected, transferred into ice-cold PBS, and then incubated in neurobasal medium (Thermo Fisher) containing 10 μg mL −1 collagenase (Sigma-Aldrich) for 15 min at 37 °C in 5% CO 2 .Media was removed and replaced with 0.5% trypsin in PBS, then incubated in a 37 °C water bath for 15 min.The solution was removed and replaced with complete neurobasal media containing 100 U mL −1 penicillinstreptomycin and 2% B-27 supplement (Thermo Fisher).Cells were gently triturated with a 1 mL pipette, filtered through a 70 μm cell strainer (Rowe Scientific, Minto, Australia) and resuspended in 7 mL of complete neurobasal culture media.The cell suspension was centrifuged at 500 g for 3 min at 4 °C, the supernatant decanted, and cells resuspended in L15 media (Thermo Fisher).TGs were separated using a Percoll gradient (Bio-Strategy, Auckland, New Zealand) after which they were placed on ɣ-irradiated lenticules and incubated in complete neurobasal media for 4 days at 37 °C in 5% CO 2 .Neurite length and density were analyzed using an automated algorithm developed by the authors, called noise-based segmentation. [75]n Vitro Stratification: Stratification was induced using transwell membranes (Corning) with 0.4 μm pores, made from polyester or polycarbonate, which were inserted into 24-well culture plates.In brief, transwell filters were rinsed with 200 μL of PBS, after which ɣ-irradiated lenticules were placed on the insert.Next, 50 μL of ≈3 × 10 5 of epithelial cells (in complete dKSFM) was dispensed onto lenticules and incubated for 30 min at 37 °C in 5% CO 2 .Then 200 μL of the same media was added to the insert, 500 μL to the bottom plate and incubated at 37 °C in 5% CO 2 .After 1 week, during which cells reached confluence, media was removed from the insert, bringing them to an air-liquid interface.Fresh complete dKSFM containing 2% FBS was added to the bottom well, and cells were cultivated for a further 2-3 weeks, exchanging the media thrice weekly.
Electron Microscopy: For TEM, cells on SMILE lenticules were prepared in fixative containing 2.5% w/v glutaraldehyde in 0.2 m sodium phosphate buffer (SPB) overnight at 4 °C, washed in 0.1 m SPB, and post-fixed in 1% osmium tetroxide (Sigma-Aldrich) in 0.2 m SPB using a BioWave Pro+ Microwave Tissue Processor (Ted Pella Inc, Redding, CA).After rinsing with 0.1 m SPB, samples were dehydrated in a graded series of ethanol (30%, 50%, 70%, 80%, 90%, and 100%), infiltrated with resin (Procure 812, Electron Microscopy Sciences, PA) and polymerized in an oven set to 60 °C for 48 h.Ultrathin sections (70 μm) were cut using a diamond knife (Diatome, Nidau, Switzerland) and collected onto carbon-coated copper slot TEM grids.Grids were post-stained using 2% uranyl acetate (Sigma-Aldrich) and lead citrate (Sigma-Aldrich).Two grids were collected from duplicate regions for each sample and imaged using a JEOL 1400 TEM (Tokyo, Japan) operating at 100 kV.For SEM, samples were fixed overnight in 2.5% w/v glutaraldehyde (Sigma-Aldrich) in 0.2 m SPB overnight at 4 °C.Fixed samples were washed three times with 0.1 m SPB followed by dehydration using a graded series of ethanol (30-100%).Next, samples were dehydrated using increasing concentrations of hexamethyldisilizane (HMDS) (Sigma-Aldrich) and air dried in 100% HMDS.Duplicate regions from each sample were mounted onto SEM stubs, platinum coated, and viewed under an FEI Nova NanoSEM 230 (Field Electron and Ion Company, OR) operating at 5 kV.
Grafting Lenticules in Injured Mouse Corneas: Intact and thinned lenticules were used in a mock transplantation experiment in mouse models of LSCD.To thin lenticules, the Schwind Amaris 750s excimer laser (Schwind Eye-Tech Solutions GmbH and Co. KG, Kleinostheim, Germany) was employed.Depths of ablation were individually calculated using the ORK-CAM (ORK-Custom Ablation Manager) software module performing a photorefractive keratectomy treatment.A 4.5 mm diameter central ablation zone was used with variable myopic treatment depending on the lenticule thickness post-SMILE surgery to equate to a thickness of ≈25 μm.Specimens were placed on the underside of a photo paper (Edmund Optics, Woodland Loop, Singapore) and kept dry prior to the ablation after which they were stored in PBS until used in experimentation.Both intact and thinned lenticules were re-sized/shaped using 1.5, 2, and/or 3 mm trephines (KAI medical, Solingen, Germany) prior to grafting.
Seven weeks-old mice were euthanized by cervical dislocation.][77][78] The right eye received two types of wounds: i) a total LSCD and ii) an annular-shaped wound.In brief, the total LSCD was created by removing the entire limbal and central corneal epithelium using an Algerbrush II attached to a 1 mm Burr.For the annular wound, the cornea was demarcated with 1.5 and 3 mm trephine and the epithelium debrided within that zone, leaving behind a central island of corneal epithelium intact.After wounding, the injured eye was copiously rinsed with saline to remove any cell debris.Next, intact and thinned lenticules were placed on injured corneas.To image the cornea and grafted lenticule, surgical microscopy (Olympus, Tokyo, Japan) and OCT (Micron IV, Phoenix Technologies, NY) were used.
Statistical Analysis: Data are presented as mean ± SD (n = sample size) and regarded statistically significant when p < 0.05.Statistical analysis was performed using Prism v9.0.2 software (GraphPad, La Jolla, CA).An unpaired t-test was used to compare the Ki67 + cells between irradiated and non-irradiated cells and to compare the length and density of neurites between Day 0 and Day 4 following cultivation of TG neurons.

Figure 1 .
Figure 1.Human and mouse corneolimbal cells cultured on human SMILE lenticules.a) Depiction of the location (arrow) from where SMILE lenticules are typically harvested.b) Lenticule transparency in saline immediately after surgery (no cells) and after seeding with hCECs for 7 days.c) Phase contrast images of hCECs cultured on a SMILE lenticule at day 7.The hatched black square is magnified to show enlarged (arrow) or detached rounded cells (arrowhead).Scale bars, 400 and 20 μm.d) A single cell suspension of hCECs was seeded on uncoated human SMILE lenticules (n = 2), and stained for pan-cytokeratin (green), p63 (red), and Hoechst (blue) at Day 7. Images were taken on a scanning confocal microscope.The hatched white square is magnified in panels i-iii which spans the anterior, intermediate, and posterior stromal layers (respectively).Scale bars, 700 and 50 μm.e) A single cell suspension of Confetti p-mCLEC was cultured on uncoated native human SMILE lenticules for 7 days (n = 7) and imaged by confocal microscopy to discern RFP fluorescent signal.The hatched white square is magnified in panels i-iii which spans the anterior, intermediate, and posterior stromal layers (respectively).Scale bars, 700 and 50 μm.

Figure 2 .
Figure 2.Human and mouse corneolimbal cells cultured on coated SMILE lenticules.a) Lenticules were coated with 10% FBS, and a single cell suspension of p-mCLECs were seeded on them for 7 days (n = 3) after which they were stained for phalloidin (red) and Hoechst (cyan).The hatched white square is magnified in panel i (right).Scale bars, 700 and 50 μm.b) Lenticules were coated with 10% FBS, and a single cell suspension of p-mCLECs was seeded on them for 7 days (n = 3) after which they were stained for K14 (red) and Hoechst (blue).The hatched white square is magnified in panel i (right).Magnified orthogonal views through the cross-section indicated by the dotted yellow line.The dashed white lines in the orthogonal views demarcate the anterior and posterior surfaces of explant (i, right).Scale bars, 700 and 50 μm.c) Lenticules cultured in dKSFM containing 10% FBS for 3 weeks (n = 3) and stained for phalloidin (green)/Hoechst (blue) or Ki67 (violet)/Hoechst (blue) (i and ii, respectively).Specimens cultured in the same media without FBS were used as control (iii) (n = 3).Representative phase contrast image of fibroblasts on expanding on tissue culture plastic which emigrated from a lenticule during 3 weeks in the dKSFM media containing 10% FBS (iv).Scale bars, 50 and 100 μm.d) Lenticules conditioned in dKSFM containing 5% FBS for 2 weeks and ɣ-irradiated at 1000cGy, cultured in FBS-free dKSFM for 2-3 days, then seeded with hCECs (left panel; n = 3).Lenticules cultured in the same condition without ɣ-irradiation were used as controls (right panel; n = 3).Cells were stained for phalloidin (green), Ki67 (purple), and Hoechst (blue).Scale bars, 40 μm.

Figure 3 .
Figure 3. Cultivation and characterization of human and mouse corneolimbal cells on lenticules using optimized culture conditions.a) Flow diagram displaying optimized lenticule preparation conditions including FBS concentration and ɣ-irradiation prior to seeding human and mouse CLECs.b,c) A single cell suspension of hCECs (n = 7) or Lifeact p-mCLECs (n = 11) were seeded on lenticules and stained for K12 (yellow), K14 (red), phalloidin (green) or GFP (green), and Hoechst (blue).Images of whole lenticules were taken by confocal microscopy (i).The hatched white square is magnified to display the basal (ii) and superficial (iii) layers respectively.Magnified orthogonal views through the cross-section indicated by the vertical and horizontal dotted yellow line.Scale bars, 700 and 40 μm.d) A single cell suspension of Lifeact p-mCLECs were seeded on lenticules (n = 5) and stained for Ki67 (red), GFP (green), and Hoechst (blue).Image of a whole lenticule was taken by confocal microscopy (i).The hatched white square is magnified in panel ii.Scale bars, 700 and 40 μm.

Figure 5 .
Figure 5. Detection of immune cells and cultivation of TG cells on lenticules.a) Immunostaining for CD45 (green), CD14 (red), III-tubulin (red), and Hoechst (blue) on lenticules immediately after receiving specimens post-SMILE surgery and after 2 weeks of pre-conditioning and ɣ-irradiation in dKSFM with 5% FBS (n = 3/group).b) Flow diagram displaying optimized lenticule preparation conditions after which murine primary TG neurons were dispensed, allowed to adhere, and then stained for III-tubulin.Confocal microscopy images were taken at day 0 and day 4 after seeding (n = 5/group).Scale bars, 70 μm.c) Nerve fiber length and density was computed at day 0 and day 4 from 5 and 11 selected images (respectively), where TG neurons were present.Bars represent mean ± SD, *p < 0.05, Unpaired t-test.

Figure 6 .
Figure 6.Lenticule resizing and reshaping for transplantation.a) 2, 3 mm, and annular shaped mini-lenticules trephined from 7 mm surgical specimens.Scale bars, 1000 μm.b,c) Mouse corneas were debrided across a 3 mm circular (b) or a 3 mm annular (c) shaped zone, and full-thickness mini-lenticules were placed over the injured mouse cornea.Phase contrast, real-time fundus view, and OCT images were taken.Scale bars, 400 and 200 μm.d,e) Mouse corneas were debrided across a 3 mm circular (d) or a 3 mm annular (e) shaped zone, and excimer ablated (25 μm thick) mini-lenticules were placed over the injured mouse cornea.Phase contrast, real-time fundus view and OCT images were taken.Scale bars, 400 and 200 μm.

Figure 7 .
Figure7.Potential SC-lenticule grafting options for patients with LSCD.We envisage that lenticules will be seeded with autologous or allogeneic LESCs.When grafts are deemed ready, patients will be scheduled into theatres to have a superficial kerectomy, that is, removal of the opaque conjunctival pannus from the cornea.After which three transplantation options could be employed.Option 1: Use several cell-loaded lenticules to cover the diseased cornea.Option 2: Use single or multiple cell-laden lenticule, cut into small wedges which are randomly distributed on the cornea.Option 3: Use single or multiple cell-laden lenticules, cut into small concentric annular-shaped strips which are place around the perimeter of the cornea.Created with BioRender.