A decellularized human corneal scaffold for anterior corneal surface reconstruction

Naresh Polisetti; Anke Schmid; Ursula Schlötzer-Schrehardt; Philip Maier; Stefan J Lang; Thorsten Steinberg; Günther Schlunck; Thomas Reinhard

doi:10.1038/s41598-021-82678-3

. 2021 Feb 4;11:2992. doi: 10.1038/s41598-021-82678-3

A decellularized human corneal scaffold for anterior corneal surface reconstruction

Naresh Polisetti ^1,^✉,^#, Anke Schmid ^1,^#, Ursula Schlötzer-Schrehardt ², Philip Maier ¹, Stefan J Lang ¹, Thorsten Steinberg ³, Günther Schlunck ^1,^✉, Thomas Reinhard ¹

PMCID: PMC7862698 PMID: 33542377

Abstract

Allogenic transplants of the cornea are prone to rejection, especially in repetitive transplantation and in scarred or highly vascularized recipient sites. Patients with these ailments would particularly benefit from the possibility to use non-immunogenic decellularized tissue scaffolds for transplantation, which may be repopulated by host cells in situ or in vitro. So, the aim of this study was to develop a fast and efficient decellularization method for creating a human corneal extracellular matrix scaffold suitable for repopulation with human cells from the corneal limbus. To decellularize human donor corneas, sodium deoxycholate, deoxyribonuclease I, and dextran were assessed to remove cells and nuclei and to control tissue swelling, respectively. We evaluated the decellularization effects on the ultrastructure, optical, mechanical, and biological properties of the human cornea. Scaffold recellularization was studied using primary human limbal epithelial cells, stromal cells, and melanocytes in vitro and a lamellar transplantation approach ex vivo. Our data strongly suggest that this approach allowed the effective removal of cellular and nuclear material in a very short period of time while preserving extracellular matrix proteins, glycosaminoglycans, tissue structure, and optical transmission properties. In vitro recellularization demonstrated good biocompatibility of the decellularized human cornea and ex vivo transplantation revealed complete epithelialization and stromal repopulation from the host tissue. Thus, the generated decellularized human corneal scaffold could be a promising biological material for anterior corneal reconstruction in the treatment of corneal defects.

Subject terms: Tissue engineering, Biomaterials - cells

Introduction

Allogenic corneal transplantation is a highly successful technique to restore vision in patients with corneal diseases or after ocular trauma¹. In eyes with moderate damage, the ocular immune privilege and the lack of corneal blood vessels decrease the risk for immunological transplant rejection. This is different in eyes with severe corneal damage, e.g. in patients suffering from alkaline burns or corneal stromal ulceration due to inflammation or infection, where allogenic transplants placed in an inflamed recipient bed are prone to transplant rejection^2,3. These patients in particular would benefit from the possibility to use non-immunogenic decellularized tissue scaffolds which may be repopulated by host cells in situ. A large number of studies used porcine tissue for decellularization^4,5. However, its clinical implementation may be hampered by issues of xenotransplantation such as immunocompatibility and the possible presence of porcine viruses. Human donor corneas unsuitable for transplantation due to a low endothelial cell count have therefore been advocated as an excellent source of the tissue^6,7. With the broad use of Descemet membrane endothelial keratoplasty (DMEK), where only a 20–30 µm thick tissue membrane and the endothelial cell layer are being transplanted, a large number of corneal tissue remnants remain after DMEK transplant preparation and could also be used for scaffold preparation to treat corneal ulcers and superficial defects.

Several methods have been described to decellularize corneal tissue using physical or chemical means^8,9. For decellularization, hypertonic sodium chloride (NaCl), sodium dodecyl sulfate (SDS) or non-ionic detergents such as Triton X-100 have been used alone or in combination with nucleases to eradicate cells while preserving protein content and structure of the extracellular matrix (ECM) scaffold to varying degrees^6,7. Drawbacks may be that SDS denatures proteins and a decrease in scaffold transparency was reported after treatment with SDS or hypertonic salt⁷. Moreover, protocols using SDS or hypertonic salt with nucleases are very time-consuming (~ 5–8 days)^6,7,10 and failure to fully elute SDS can have cytotoxic effects^6,7. Sodium deoxycholate (SD), a naturally occurring bile salt metabolite, has been used to decellularize various organs or tissues such as liver¹¹, peripheral nerves¹², ovary¹³, uterus¹⁴, conchal cartilage¹⁵, trachea¹⁶, lungs and heart valves¹⁷. The combination of SD/ deoxyribonuclease I (DNAse) was also used to generate decellularized human corneal limbus, however, a long elution time for removal of cellular components was a drawback of the protocol suggested¹⁸. Moreover, there are no reports on a combination of SD and DNAse as a decellularizing agent for human corneal tissue. Synthetic SD is being applied clinically as an injectable dissolvent of subcutaneous body fat¹⁹ and was FDA-approved for removal of submental fat, which may illustrate its biocompatibility. The cornea is endowed with high glycosaminoglycan content and it has been shown that tissue swelling during processing can cause structural damage, which was prevented by adding dextran to the solutions used on porcine corneal tissue²⁰. The role of dextran and tissue swelling during human corneal decellularization has not specifically been addressed in earlier protocols^6,7. Moreover, published decellularization protocols often lack an extensive characterization of ECM integrity and preservation of BM components^6,7, which are essential to maintain tissue homeostasis and ensure cell adhesion, growth, long term survival and function of corneal epithelium and epithelial progenitor cell phenotype^8,21,22. SDS and triton-x-100 were shown to be cytotoxic and limit biocompatibility^7,23.

In light of these data, it was our goal to devise a further improved protocol for decellularization of human corneal tissue remnants after removal of transplant material for DMEK (without Descemet membrane, labeled DM⁻) using agents in clinical use and controlling for tissue swelling. We also wanted to clarify whether corneal tissue unsuitable for transplantation due to a low endothelial cell count (with Descemet membrane, labeled DM⁺), would be amenable to the same protocol with similar preservation of tissue integrity. We evaluated decellularization effects on the basement membrane (BM) components and physical characteristics such as transparency and mechanical properties of the decellularized human cornea (DHC). Furthermore, we evaluated the biocompatibility of generated DHC scaffolds by repopulation with human corneal cells from the limbus.

Results

Decellularization of human corneas

Human DM⁻ corneas decellularized with different concentrations of SD were initially screened by hematoxylin and eosin (H&E) staining, immunofluorescent staining for human leukocyte antigen (HLA)-ABC was used to assess the presence of residual cellular/membranous material, and nuclear staining with 4′,6-diamidino-2-phenylindole (DAPI) (Fig. 1A). H&E staining showed distinctly visible corneal epithelial and stromal cells (arrowheads indicating stromal cells at higher magnification) in normal human cornea (NHC) (Fig. 1A). In DHC treated with 4% SD or 1% SD, no epithelial and stromal cells were observed and HLA-ABC, as well as deoxyribonucleic acid (DNA) (DAPI), was also not detected (Fig. 1A). The use of 0.5% SD did not result in complete removal of cellular debris as apparent in H&E (Fig. 1A, higher magnification, arrowheads) and HLA-ABC staining’s (Fig. 1A, arrows). These results were reproduced in DM⁺ corneas (data not shown). Thus, a concentration of 1% SD was used for decellularization in all subsequent experiments. We did observe an increased thickness of DHC as compared to the respective NHC controls. Final swelling states of DHC in the DM⁻ and DM⁺ groups were also different depending on the swelling state at the outset of the decellularization procedure (Table 1).

Decellularization efficiency and role of dextran: (A) Histological evaluation of human cornea decellularized by various concentrations of sodium deoxycholate (SD) with deoxyribonuclease I in comparison to the normal human cornea (NHC) using hematoxylin and eosin (H&E) staining. The higher magnification images of H&E staining showing the stromal cells in NHC (arrowheads) and cellular debris in 0.5% SD ( arrowheads). Scale bar = 100 µm. HLA-ABC staining (green) of NHC and DHC showing cellular expression of HLA in NHC and 0.5% SD treated DHC (white arrows); no HLA expression in 1% and 4% SD treated DHC cornea. DAPI staining reveals nuclei in NHC without remaining nuclei/nuclear debris in DHC. (B) The role of dextran during decellularization in both DM⁻ and DM⁺ cornea in comparison to respective controls. Graphs represent the thickness of DHC in various dextran conditions compared to NHC. Data are expressed as means ± S.E.M. (n = 5). * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001; (Wilcoxon signed-rank test). (C) AS-OCT images of DM⁻ and DM⁺ cornea in varying concentrations of dextran with respective controls. (D) Histological evaluation by staining with H&E on NHC and DHC showing no cell remnants in DHC but varying corneal thickness (in presence or absence of dextran) of both DM⁻ and DM⁺ cornea. Scale bar = 100 µm E) Confirmation of decellularization by quantification of residual DNA. The graph represents the quantity of DNA in both DM⁻ and DM⁺ cornea before and after decellularization (0 or 4% dextran). Data are expressed as means ± S.E.M. (n = 4). * p < 0.05; Mann–Whitney U test. *DAPI* 4′,6‐diamidino‐2‐phenylindole, DM⁻ cornea with the absence of Descemet membrane, DM⁺ cornea with the presence of Descemet membrane, *AS-OCT* anterior segment optical coherence tomography, *HLA* human leukocyte antigen.

Table 1.

Corneal thickness measurements by anterior segment optical coherence tomography.

Corneal tissue	Treatment	Dextran during treatment (%)	Thickness (µm) after treatment (mean ± S.D.)	Thickness (µm) after additional 24 h in 6% dextran (mean ± S.D.)
DM⁻	NHC	0	1243.4 ± 108.3
	Dextran Control	6	702.6 ± 57.4 (24 h)	548.0 ± 19.2 (48 h)
	DHC	0	1487.6 ± 103.2	686.3 ± 35.0
		2	1169.9 ± 71.7	609.0 ± 11.9
		4	1043.7 ± 112.7	536.6 ± 24.7
		6	1005.8 ± 96.2	540.1 ± 30.5
DM⁺	NHC	6	554.1 ± 55.4
	Dextran-free Control	0	711.2 ± 86.6 (24 h)	528.0 ± 31.9
	DHC	0	1192.1 ± 125.2	477.0 ± 37.5
		2	902.2 ± 95.4	581.3 ± 36.2
		4	679.2 ± 59.7	536.3 ± 9.2
		6	649.4 ± 76.9	593.6 ± 10.5

Open in a new tab

Effects of dextran during decellularization

In DM⁻ corneas, the average thickness of control NHC was 1243.4 ± 108.3 µm, whereas DHC processed in the absence of dextran (1487.6 ± 103.2 µm, p = 0.03) were significantly thicker (Fig. 1B). Decellularization in 2% dextran yielded thinner scaffolds (1169.9 ± 71.7 µm), in 4% dextran (1043.7 ± 112.7 µm, p = 0.007) and 6% dextran (1005.8 ± 96.2 µm, p = 007) the reduction in corneal thickness was significant when compared to (NHC) (Fig. 1B). As dextran control, the NHC were incubated in a medium containing 6% dextran with an incubation time equivalent to the decellularization process (~ 24 h) resulted in a significant decrease in thickness (702.6 ± 57.4 µm, p = 0.0001) (Table 1).

In DM⁺ cornea, the average thickness of NHC was 554.1 ± 55.4 µm, whereas DHC were swollen in the absence of dextran (1192.1 ± 125.2 µm, p = 0.0001) and less so in 2% dextran (902.2 ± 95.4 µm, p = 0.003, Fig. 1B). The mean thickness of the scaffolds decellularized using 4% dextran (679.2 ± 59.7 µm) and 6% dextran (649.4 ± 76.9 µm) was best comparable to NHC (Fig. 1B). As dextran-free control, the NHC were incubated in a medium containing 0% dextran with an incubation time equivalent to the decellularization process (~ 24 h) resulted in a significant increase in thickness (711.2 ± 86.6 µm, p = 0.0001) (Table 1).

The corresponding anterior segment optical coherence tomography (AS-OCT) images also reflected the same (Fig. 1C). When corneas (both DM⁻ and DM⁺) were kept at 6% dextran for 24 h after decellularization, corneal thickness was comparable to NHC irrespective of the dextran concentration used in decellularization (Table 1). Based on these results, we performed all subsequent experiments using 1% SD with 4% dextran during the whole decellularization process, compared to 1% SD without dextran. Even though the 2% dextran was effective to reduce the thickness of DM⁻ corneas (1169.9 ± 71.7 µm) to match NHC (1243.4 ± 108.3 µm), we used 4% dextran for all further experiments as it was our goal to devise a protocol amenable to both DM⁻ and DM⁺ corneas.

To test whether the decellularization in the presence of 4% dextran is as efficient as in its absence, histological evaluations were performed in both DM⁻ and DM⁺ corneas. H&E staining showed a decrease of corneal thickness when dextran was present during decellularization, but no differences were noticed with respect to remnant cellular material. This was true in DM⁻ as well as DM⁺ samples (Fig. 1D). Spectroscopic analyses of DNA content showed significantly less residual DNA in DHC compared to NHC, irrespective of dextran used during decellularization. SD and DNAse treatment removed about 98.6 ± 1.2% of the DNA content in the presence of dextran and 98.4 ± 0.9% in its absence (Fig. 1E). We also noticed that the quality of DM⁺ corneas was poor compared to DM⁻ (loss of epithelial layers in (Fig. 1D), and showed lower DNA content (Fig. 1E) presumably due to long time incubation in dextran containing medium as mentioned earlier (Table 2)²⁴.

Table 2.

Organ cultured corneal sample details.

Corneal samples	Number	Donor age (years)*	Cultivation duration (days)*	Cultivation in dextran (days)*
Total	183	70.2 ± 13.0 (24–93)	32.0 ± 9.5 (6–75)	4.4 ± 6.5 (0–43)
DM⁻	102	69.4 ± 10.3(42–87)	30.0 ± 6.4 (16–58)	1.4 ± 1.1 (0–3)
DM⁺	51	77.3 ± 11.4 (50–93)	35.3 ± 11.7 (6–69)	12.5 ± 8.4 (3–43)
After PK	30	61.3 ± 16.7 (24–86)	35.0 ± 14.6 (20–75)	2.3 ± 1.1 (1–4)

Open in a new tab

*All values are expressed as mean ± standard deviation (range).

Scanning electron microscopy

The effect of decellularization treatment on the integrity of the ECM structure was evaluated by scanning electron microscopy (SEM). As dextran control, the NHC were incubated in a medium containing 4% dextran with an incubation time equivalent to the decellularization process. Epithelial cells detectable in control corneas (Fig. 2A, i & iii) were absent in DHC (Fig. 2A, ii & iv), irrespective of dextran use. On cross-sections, no significant gross changes of the collagen fibers were noted in DHC except that the collagen bundles were slightly larger due to swelling and a slight increase in distance between bundles in absence of dextran (Fig. 2A, vi). When decellularized in the presence of 4% dextran, DHC showed a decrease in collagen bundles distance (Fig. 2A, viii) compared to NHC in dextran-free medium (Fig. 2A, v).

Optical properties

Macroscopic tissue evaluation suggested that decellularization in the presence of dextran did not affect transparency or even slightly improved it, whereas decellularization in the absence of dextran led to some tissue clouding (Fig. 2B, 0 h). Twenty-four hours of incubation in glycerol resulted in the recovery of DHC transparency in the absence of dextran (Fig. 2B, 24 h). The spectroscopic analysis at 300–750 nm revealed similar transparency of NHC and DHC processed in 4% dextran (Fig. 2C). In the absence of dextran, there was a reduction in transparency by 10–15% (Fig. 2C, 0 h), which subsided after 24 h in glycerol (Fig. 2C, 24 h).

Mechanical properties

Instrumented indentation measurements indicated no major differences in the elastic moduli of NHC (62.6 ± 15.9 kPa) and DHC irrespective of dextran (59.3 ± 13.6 kPa in 0% dextran; 59.2 ± 10.56 kPa in 4% dextran) used during decellularization (Fig. 2D).

ECM components

In periodic acid schiff (PAS) staining, the intensity of the glycoproteins of DHC seemed markedly reduced in the absence of dextran as the thickness of the cornea also increased (Fig. 2E) whereas the intensity was increased in the DHC in the presence of dextran (Fig. 2E), compared to NHC as thickness also decreased (Fig. 2E).

Alcian Blue (AB) staining of NHC showed the presence of glycosaminoglycan (GAG) throughout the cornea. In DHC, GAG staining intensity was much weaker when treated in the absence of dextran (Fig. 2E), whereas in DHC treated in the presence of dextran GAG staining was more intense than in NHC (Fig. 2E). The levels of sulfated GAG (sGAG) measured with biochemical assays were unaffected by decellularization and no significant differences were observed among the samples (Fig. 2F).

The ECM of the corneal surface was evaluated in detail by immunostaining for various components including agrin, collagen types (Col III, IV, XVIII), fibronectin (FN), junctional adhesion molecule C (JAM-C), perlecan, tenascin C (TN-C), vitronectin (VN), and laminin (LN) chains (α3, α5, β2, β3, and γ2).

The major BM heparan sulfate proteoglycan, agrin, was expressed in BM with uniform and strong labeling intensity in all samples (Fig. 3, arrowheads). Expression of Coll IV and Coll XVIII molecules were clearly noted in BM of the cornea in both NHC and DHC samples (Fig. 3, arrowheads). No significant difference was noted for Col IV, whereas Coll XVIII expression was reduced in DHC (no dextran) and discontinuous in DHC (4% dextran) compared to NHC. The fibrillar Coll III was not observed in either NHC or DHC (Fig. 3).

Extracellular matrix (ECM) composition of decellularized human cornea (DHC). Immunostaining analysis of various (ECM) and basement membrane (BM) molecules on the DHC (0 and 4% dextran) compared to the normal human cornea (NHC). The white dashed line marked the boundary between the limbal and corneal tissue in laminin (LN) α5 staining of NHC. Arrowheads indicating the expression of ECM and BM molecules. Nuclei counterstained with 4′,6‐diamidino‐2‐phenylindole (blue). *Col* Collagen, *JAM-C* junctional adhesion molecule C.

FN was clearly detected in BM in all samples without any significant differences (arrowheads), whereas a weak expression of vitronectin was observed (Fig. 3). Strong expression of the other major BM heparan sulfate proteoglycan, perlecan, was noted similar to agrin (Fig. 3, arrowheads). TN-C was hardly expressed on the BM of the corneal surface of all samples, whereas JAM-C expression was not observed (Fig. 3). Using Laminin chain specific antibodies, the LN-α3, -ß3, -γ2 chains were found to be strongly expressed in the corneal BM of NHC and DHC with or without dextran (Fig. 3, arrowheads). No significant differences were observed in staining patterns among samples. The LN α5 chain and β2 chains were not detectable in corneal samples as their expression is known to be limited to the limbal region (Fig. 3)²⁵.

Recellularization

The characterization details of cultured limbal epithelial progenitor cells (LEPC), limbal mesenchymal stromal cells (LMSC), and limbal melanocytes (LM) were provided in the Supplementary Data S1.

In vitro recellularization

The biocompatibility of the DHC to support LEPC, LMSC, and LM was evaluated. The seeded LEPC and LMs attached to the surface of decellularized corneas (Fig. 4A, i-iii) and remained viable after 48 h of cultivation as confirmed by live/ dead staining (Fig. 4A, iv&v). In DHC-LEPC/LM scaffolds, the melanocytes intermingled with epithelial cells (Fig. 4A, iii, white asterisk). The LMSCs injected into DHC spread and appeared healthy (Fig. 4A, vi). H&E staining showed a monolayer of LEPC on DHC after 1 week of cultivation (Fig. 4B, 1 w). After 3 weeks in culture stratification of the epithelium was successfully induced by lifting the tissues to the air–liquid interface and confirmed by H&E staining (Fig. 4B, 3 w). The injected LMSCs were spread over the posterior side of the DHC after 1 week (Fig. 4C, 1 w). After 3 weeks, the LMSCs appeared to migrate into the anterior part of the corneal stroma (Fig. 4C, 3 w). After 3 weeks of cultivation, the DHC-LEPC/LM scaffolds showed a stratified surface epithelium and stromal cells in the anterior part of the stroma as confirmed by H&E staining (Fig. 4D). The DHC-LEPC/LM scaffolds also showed a stratified epithelium after 3 weeks of cultivation (Fig. 4E).

Phenotypic evaluation of recellularized scaffolds by immunofluorescent staining confirmed pronounced pan-cytokeratin (pan-CK) and E(epithelial)-cadherin staining in all epithelial layers; CK15 and p63 staining confined to basal layers (Fig. 4F, arrowheads); corneal differentiation marker CK3 was expressed in superficial epithelial layers but not in basal cells (Fig. 4F, arrows indicating basal cells) in DHC-LEPC scaffolds . Vimentin-positive cells (red) were observed in the stromal region close to pan-CK-positive epithelial cells (green) of the DHC-LEPC/LMSC scaffolds (Fig. 4G). Melan-A positive cells (red) were interspersed in the epithelial layers (pan-CK⁺, green) of the DHC-LEPC/LM scaffolds (Fig. 4H, arrowhead). The results of recellularization by neighboring tissue are provided in Supplementary Data S2.

Ex vivo transplantation

The anterior part of a DHC was successfully sutured onto the posterior bed of a NHC (Fig. 5A) and both tissues were well adapted by sutures as confirmed by AS-OCT imaging (Fig. 5B). The surgical handling of the DHC tissue was comparable to NHC and no significant tissue damage occurred during suturing. After 2 weeks in culture, histological analyses showed complete epithelialization of the grafted tissue scaffold with stratification and the absence of cells in the stromal region of the graft (Fig. 5C, 2 w). Stromal cells were detectable in the graft 5 weeks after surgery (Fig. 5C, 5 w). Immunohistochemical analyses revealed the expression of the epithelial markers cytokeratin (pan-CK) and E-cadherin in the epithelial cells on the host (peripheral cornea/limbal region) and graft tissue (Fig. 5D). Expression of the LEPC markers, CK15, P(placental)-cadherin, and p63, was apparent in basal epithelial cells on both, the grafted scaffold and host tissue, whereas expression of the corneal differentiation marker CK3 was present only in the superficial epithelial cells of the graft and host tissue (Fig. 5D). Moreover, we also observed Ki-67 positive cells in the basal layer of the epithelium covering the DHC tissue (Fig. 5D, arrowheads). The migrated stromal cells observed in the graft tissue after 5 weeks showed expression of vimentin and keratocan, a corneal stromal marker (Fig. 5D, arrowheads). Interestingly, we observed Melan-A positive melanocytes in association with basal epithelial cells in the graft tissue similar to host tissue (Fig. 5D, arrowheads).

Discussion

Decellularized corneal scaffolds recently gained interest as an alternative material for corneal replacement^26–28. If corneal damage extends from the ocular surface to the upper stroma and an allotransplant is at high risk of rejection due to vascularization or inflammation, decellularized corneal scaffolds may serve as an alternative tissue source to reconstruct the corneal epithelium and its underlying stromal support. Therefore, it was our goal to devise an efficient, fast, and gentle method of decellularization using well-tolerated substances to generate human corneal scaffolds from readily available donor material.

Various protocols have been used to decellularize human corneal tissue, of them, SDS or hypertonic salt with nucleases were most effective in removing cellular components^6,7. The combination of SD, a natural bile salt, and DNAse was used to decellularize human corneal limbus. However, significant drawbacks were the long elution times required for removal of cellular components by SDS/nucleases or NaCl/nucleases treatment reported for human corneal decellularization (~ 5–8 days)^6,7 and SD/DNAse for treatment of human limbal tissue (~ 5 days)¹⁸. In this study, we established and validated a rapid alternative method (~ 1 day) for efficient decellularization of the human cornea using SD and DNAse. Our data indicate that 1% or 4% SD with DNAse and 4% dextran allow for efficient decellularization by removal of cellular components (HLA-ABC and nuclear debris) within 24hrs, in line with earlier reports on other tissues^11,17,18.

Due to its high glycosaminoglycan content, corneal tissue has a strong tendency to swell in cell culture media^29–31. A concentration of 5–6% dextran in a medium is therefore routinely used to maintain graft thickness prior to keratoplasty^31,32. The significant difference of the mean NHC thickness between DM⁻ and DM⁺ groups is due to the presence or absence of dextran in the respective culture medium used for tissue transfer from the Cornea Bank to the laboratory. Most previous reports on human corneal decellularization have not specifically addressed tissue swelling^6,7, but dextran use has been reported on porcine corneas in conjunction with ultrahigh hydrostatic pressure and chemical detergents^20,29,30. We compared different concentrations of dextran to control tissue swelling during decellularization of the human cornea and found that the presence (DM⁺) or absence (DM⁻) of Descemet membrane had an additional influence on tissue thickness (Fig. 1B–D). In both, DM⁺ and DM⁻ cornea, a dextran concentration of 4% was sufficient to preserve corneal thickness during decellularization. In line with earlier reports on porcine tissue^7,20,29,30, SD/DNAse-mediated removal of cellular components and DNA from human corneal tissue were unhampered by the presence of dextran (Fig. 1D,E). Moreover, our decellularization protocol is also applicable to DM⁺ corneas, which might be usable for full-thickness corneal transplantation in the future.

As optical properties are highly relevant for the selection of materials for corneal reconstruction, we next assessed transparency and optical transmittance of DHC. It has been reported that ionic detergents solubilize both cytoplasmic and nuclear cell membranes, damage collagen, and denature proteins by disrupting protein–protein interactions^33,34. In a previous study, SDS/DNAse-treated corneas had the greatest loss of transparency with most ECM disruption as compared to corneas treated with non-ionic detergents⁷. Even though SD is an ionic detergent, ECM disruption, or reduction of the sGAGs content were not observed similar to previous studies on other tissues³⁵. In our own experiments, the transparency of DHC decreased with decellularization in the absence of dextran and recovered after incubation in glycerol for 24 h (Fig. 2B). This suggests that the decrease in corneal transparency was a result of increased corneal hydration rather than permanent structural alteration^6,29. However, it has also been reported that glycerol causes additional collagen disorganization³⁶ and may nevertheless mask structural damage inflicted by the decellularization procedure⁷. Herein, we suggest to include dextran during decellularization of the human cornea to reduce tissue swelling, loss of transparency, and to maintain tissue integrity.

Ex vivo cultivated limbal epithelial cells have been used to reconstruct the corneal epithelium in limbal stem cell-deficient patients^37–40. The long term clinical outcome of limbal epithelial cell grafts mainly depends on the maintenance of stem/progenitor cells, which is regulated by the niche microenvironment including its ECM components and mechanical properties in vivo^39,41–44. Moreover, ECM molecules play a key role in cell–cell interactions, cell adhesion, proliferation, migration, and response^25,45. Thus, the integrity and preservation of ECM components are essential for further clinical use of the decellularized tissues. The published decellularization protocols have not reported on ECM component preservation in much detail^6,7. To address this issue, we extensively studied ECM components and their possible alteration. The corneal BM collagens type IV, important in the structure and function of BM, and type XVIII, having an anti-angiogenic activity^46,47, were well preserved by our decellularization protocol. Heparin sulfate proteoglycans (e.g. perlecan and agrin) and glycoproteins (e.g. fibronectin), which were well preserved in this study, play important roles in epithelial cell migration and proliferation⁴⁸. LNs are among the best described BM components of the cornea, of which the cornea-specific isoform LN-332 promotes adhesion, migration, and differentiation of LEPC^25,49. Our own data clearly indicate the preservation of a vast panel of ECM proteins including various laminin isoforms as detected by immunofluorescence microscopy (Fig. 3).

Successful repopulation with cells is a prerequisite for the use of decellularized scaffolds in tissue engineering. It has been reported that structurally damaged decellularized cornea and residual decellularizing agents (SDS) detrimentally affect cell growth and proliferation⁷. It has been shown that SD-produced scaffolds were highly biocompatible compared to those decellularized using SDS²³. We explored different strategies of DHC repopulation by plating cells or by allowing cellular invasion from host tissue. With both approaches, the repopulating epithelial cells maintained a proper phenotype as characterized by expression of the progenitor markers CK15, P-cadherin, and p63 in basal layers and the marker for differentiated epithelial cells, CK3, in the superficial layers. Stromal repopulation is a slower process as has been clinically observed in keratoconus patients following collagen cross-linking treatment. Activated keratocytes started to populate the corneal stroma at 2 months after treatment and repopulation was almost complete after 6 months⁵⁰. We observed stromal repopulation of a scaffold by the 3rd week in vitro, whereas it took 5 weeks in an ex vivo model. Thus, the repopulation of DHC by stromal cells could find an application treating anterior corneal lesions where corneal stroma is altered by disease, injury, or scarring⁵¹. Currently, no data are available regarding the repopulation of decellularized tissue by LMs, which are found in close proximity to LEPC at the limbus in vivo, where melanocytes appear to play a supporting role in preserving the stemness state of corneal epithelial cells and modulating their migration towards the central cornea during regeneration^52,53. LMs were able to adhere and intermingle with epithelial cells on DHC tissue, strongly suggesting the potential of the scaffold to facilitate LM survival. Moreover, in ex vivo lamellar grafting experiments we have also observed the migration of LMs from donor limbal tissue onto decellularized corneal grafts similar to reports on ocular surface regeneration in vivo^52,54. All these observations indicate that the DHC scaffold provides a niche microenvironment to keep basal epithelial cells in a less differentiated state. Thus, it may serve as an ideal scaffold for future clinical transplantation purposes^6,45. However, in vivo studies are necessary to determine whether the DHC supports stromal dehydration, long-term transparency, and in vivo biocompatibility. Moreover, it has to be tested whether a significant reduction of immunogenic epitopes was achieved to allow for long-term immunological tolerance.

In conclusion, we have developed a fast and efficient corneal decellularization method using clinically applicable SD with DNAse in the presence of dextran. Dextran prevented corneal swelling during decellularization. The biomechanical and optical properties, ECM architecture and BM composition of the DHC were well preserved and resembled the properties of cultured NHC. The DHC scaffolds revealed excellent biocompatibility and ex vivo transplanted corneas were completely epithelialized and effectively supported stromal cell ingrowth. Hence, this could be a promising scaffold for anterior corneal reconstruction in the treatment of corneal defects.

Methods

Tissue

Organ cultured human corneoscleral tissue remnants after DMEK (DM⁻, n = 102) and corneal buttons unsuitable for transplantation (DM⁺, n = 51) were used (Table 2). To isolate primary limbal cells for repopulation experiments, organ cultured corneoscleral rims remaining after penetrating keratoplasty were used (labeled after PK, n = 30, Table 2). The tissue was kindly provided by the LIONS Cornea Bank Baden-Württemberg, located at the Eye Center, University of Freiburg and informed consent for research use of remnant tissue had been given by the donors or their next of kin. The study was approved by the institutional review board of the University of Freiburg (25/20) and followed the tenets of the Declaration of Helsinki. No organs or tissues from prisoners were used.

Following the standard procedures for corneal grafts in the Eye Center, we obtained DM⁺ corneas from the Cornea Bank cultured (12.5 ± 8.4 days) in Carry-C medium (Alchimia) containing 6% dextran, whereas DM⁻ corneas were cultured also in Carry-C medium (1.4 ± 1.1 days) before DMEK surgery and were supplied to the laboratory in Tissue-C medium (Alchimia) without dextran. Due to higher availability, most experiments were performed on DM⁻ corneas except as otherwise indicated. DM⁺ corneas were especially used to comparatively analyze the decellularization protocols in the course of initial experiments.

Decellularization

Corneas without any further processing were used as controls (NHC). For decellularization, all washing steps and SD (Sigma-Aldrich, Hamburg, Germany) incubation were carried out under continuous agitation (800 rpm) at room temperature. Corneoscleral tissue was washed in Dulbecco’s phosphate-buffered saline (DPBS; 3 × 5 min), placed in 12-well plates with 0.5%, 1%, or 4% SD in ultrapure water for 30 min, and then rinsed in DPBS (3 × 30 min). Subsequently, the tissues were incubated in DNAse I , 1 mg/ml in DPBS (Roche, Mannheim, Germany) overnight under a sterile hood and terminally washed in DPBS (4 × 30 min). The whole decellularization procedure was carried out under aseptic conditions to ensure tissue sterility.

Quantification of tissue swelling

To assess the effects of dextran on corneal swelling during decellularization, all decellularization steps were performed in the presence of 0, 2, 4, or 6% dextran 500 (Sigma-Aldrich, Hamburg, Germany) in both DM⁺ and DM⁻ corneas. To verify the effect of dextran alone on NHC, DM⁻ corneas in incubated in 6% dextran (dextran control) and DM⁺ corneas in dextran free medium (dextran-free control) with an incubation time equivalent to the decellularization process. Corneal swelling was analyzed by AS-OCT using a clinical AS-OCT device (SS-1000 CASIA, Tomey, Nuernberg, Germany). Measurements of the central corneal thickness were performed with 3-dimensional cross-sectional imaging before and after decellularization as described earlier⁵⁵.

Histology

For bright field light microscopy, tissue was cryosectioned and stained as follows: After embedding and freezing in optimal cutting temperature medium, 10 µm thick sections were cut with a cryostat (Leica, Germany), mounted on adhesive slides and air-dried. Sections were stained with hematoxylin (Haematoxylin Gill III, Surgipath, Leica, Germany) for 2 min and 1% eosin Y (Surgipath, Leica, Germany) for 1 min to observe the gross tissue architecture and degree of decellularization. To visualize the proteoglycan content, cryosections were stained with 1% alcian blue (Morphisto, Germany) for 30 min together with a counterstain of nuclear fast red (Morhphisto, Germany). For glycoproteins, PAS staining was performed using 1% periodic acid (Fluka, Germany) for 10 min, and Schiff reagent (Roth, Germany) for 90 s. Samples were examined using either a bright field fluorescence microscope (Axio Imager.A1, Zeiss) and images were processed using ProgRes CapturePro Software (JENOPTIK, https://www.jenoptik.com/products/cameras-and-imaging-modules/microscope-cameras/progres-usb-20-firewire/software-download-progres) or a Hamamatsu NanoZoomer S60 (Hamamatsu Photonics, Herrsching, Germany).

Immunostaining of frozen sections

Immunostaining on frozen sections performed as previously described⁶⁰. Briefly, human corneas in optimal cutting temperature medium were cut into 10 µm sections, fixed in 4% paraformaldehyde (PFA) for 20 min or acetone for 10 min followed by permeabilization in 0.3% Triton X-100 in PBS for 10 min. The sections were blocked with 10% normal goat serum (NGS) and incubated in primary antibodies (Supplementary Table S1) diluted in 1% NGS in PBS overnight at 4 °C. Fluorescein isothiocyanate-conjugated or rhodamine-conjugated anti-mouse or -rabbit immunoglobulins (Life Technologies, Carlsbad, CA) were used for detection and nuclear staining was performed with DAPI (Vectashield antifade mounting medium with DAPI; Vector, Burlingame CA). Immunolabeled cryosections were examined with a laser scanning confocal microscope (TCS SP-8, Leica, Wetzlar, Germany). For negative controls, the primary antibodies were replaced by equimolar concentrations of an irrelevant isotypic primary antibody of the same species.