The human corneal stroma is responsible for two-thirds of the refractive power of the eye and occupies 90% of the corneal thickness. When affected by disease or trauma, the homeostasis and thus transparency of the tissue is compromised. This is partially due to presence of local edema and activation of resident corneal cells—keratocytes. These quiescent cells assume a dendritic cell morphology
Cells derived from the human corneal stroma have been shown to possess trilineage differentiation potential and presence or absence of specific markers (e.g., CD73+, CD90+, CD105+, CD34−, CD45−, CD14−, CD11b−, CD79
Establishment of CSC cultures is relatively easy. Nevertheless, many countries have access only to peripheral tissue remaining in unused corneal rings after keratoplasty. Use of enzymatic digestion to isolate CSCs out of the tightly packed collagen layers appears straightforward as well [
Various types of culture media have also been assessed to elucidate the best possible conditions for induction of
Cadaveric tissue collection complied with the directive of the Helsinki Declaration and was approved by the National Medical Research Council (14387/2013/EKU-182/2013). Samples were obtained within 24 hours from death. Following disinfection by povidone iodine (Egis, Hungary) and rinsing with PBS of human bulbi, corneal buttons were dissected using scissors. The corneal epithelium, Descemet’s membrane, and corneal endothelium were peeled off. To obtain equal-sized stromal explants, pieces of tissue measuring 3 mm in diameter were punched out from the corneal buttons with the help of a trephine from regions defined as the peripheral versus the central stroma, as shown in Figure
Anatomical features of the human cornea and sites of stromal cell isolation. Pieces of tissue were punched out from the indicated central and peripheral corneal regions by a surgical trephine (3 mm) for consequent digestion and/or culturing. The picture was taken by a phase contrast microscope and put together as an overlay (EVOS® FL microscope, Thermo Fisher Scientific).
Four types of CSC cultures were defined as central explant (CE) and peripheral explant (PE) versus central digested (CD) and peripheral digested (PD) from the same donors, referring to the location and presence or absence of enzymatic digestion, respectively. 24-well plates (Corning Costar, Sigma-Aldrich) were used to expand the cells. Dulbecco’s Modified Eagle Medium (DMEM) (Sigma-Aldrich) supplemented with 10% Fetal bovine serum (FBS) (Sigma-Aldrich) and 1% penicillin-streptomycin (PS) (Sigma-Aldrich) was applied to the cells. Culture media was changed every alternate day. Cells up to passage 4 were used for the experiments.
CD, PD, CE, and PE were expanded in 24-well culture plates (Corning Costar). Cells were fixed in 4% paraformaldehyde (Sigma-Aldrich) and permeabilised using 0.1% Triton X-100 (Sigma-Aldrich). Bovine serum albumin (BSA) 1% (Sigma-Aldrich) diluted in phosphate-buffered saline (PBS) was applied as a blocking solution for 1 hour at room temperature. Samples were incubated with the primary Ki-67 (Sigma-Aldrich) antibody for 1 hour at room temperature. A phycoerythrin-conjugated secondary antibody was used to visualize the protein and finally, 4′,6-diamidino-2-phenylindole (DAPI, Sigma-Aldrich) counterstaining to stain the cell nuclei. Pictures were taken by an EVOS FL microscope (Thermo Fisher Scientific).
For studying the protein expression in situ, corneal sections were prepared from paraffin-embedded tissues and stained for markers expressed by progenitor and/or stem cells (ABCG2, CXCR4, and Nestin). Proliferation- (Ki-67), function-related (ALDH1A1, Collagen I, and CD34), and MSC markers (CD73, CD90, CD105, and Vimentin), extracellular matrix and cell-adhesion components (Fibronectin, Collagen IV, and VE-Cadherin), and other molecules (
In brief, sections were deparaffinised and the nonspecific sites were blocked by 1% BSA (Sigma-Aldrich) for 1 hour at room temperature. Primary antibodies were applied overnight at 4°C. Following three times 5-minute wash by PBS containing 1% Tween-20 (PBST), Alexa Fluor 488 conjugated secondary antibodies were incubated on the sections for 1 hour at room temperature. DAPI counterstaining was performed to visualize the nuclei. Pictures were taken by a Zeiss Axio Observer Z1 (Carl Zeiss) microscope.
CSCs were subcultured in 150 cm2 flasks (TPP, Sigma Aldrich) for FACS analyses. Cells were collected by trypsinisation (Hyclone, GE Healthcare Life Sciences, Logan, Utah, USA) for surface protein expression analyses. After centrifugation at 1000 RPM, for 10 minutes, the cells were resuspended in FACS buffer (0.05% Na-azide and 0.5% BSA in DPBS). Three-color staining—fluorescein-isothiocyanate, phycoerythrin, and allophycocyanin-conjugated primary antibodies against ABCG2, CD31, CD34, CD44, CD47, CD49a, CD49d, CD51, CD73, CD90, CD105, and Nestin—were applied for 30 minutes at 4°C. FACS Calibur cytometer (BD Biosciences, Immunocytometry Systems) was used to measure the samples. Finally, data were analysed by Flowing Software 2.5 (Perttu Terho, Turku Centre for Biotechnology, University of Turku, Finland) and FCS Express 6 (De Novo Software, California, USA). More information about the antibodies is provided in Table S
Native corneal stroma tissue free of epithelium and endothelium was collected from 3 donors. The tissue from the donors was homogenized and pooled. Total RNA was isolated by Qiazol reagent (Qiagen) and RNeasy mini kit (Qiagen) following the manufacturer’s protocol. Similarly, total RNA was isolated from cultured cells separately for the 4 defined conditions by the RNeasy mini kit and pooled from 3 donors.
Nucleic acid concentrations were determined by a NanoDrop spectrophotometer (Thermo Fisher Scientific). Random hexamers and Superscript III reverse transcriptase (Life Technologies, Waltham, MA, USA) were used to transcribe 1
One-way ANOVA and Student’s
Cultures established by the enzymatic method (CD, PD) yielded CSCs immediately, and these cells proliferated fast (Figure
Phase contrast images of corneal stroma cell cultures and respective growth rates. Pictures show days 1, 8, 15, and 21 of cultivation for cells obtained from the central and peripheral regions of the stroma by enzymatic digestion and explant techniques (a). The scale bars represent 1000
Ki-67 staining revealed a strong proliferative capacity of the cells in both the explant and the digested cultures isolated from the central and peripheral regions of the corneal stroma (Figure
Immunofluorescent staining of nuclear Ki-67 in cultured cells. CSCs obtained from either the central or peripheral corneal stroma by digestion and explant techniques have been cultured for 21–30 days, respectively, and stained for proliferation marker Ki-67. The proliferation marker is shown in red with DAPI counterstaining (a). Relative quantity of Ki-67-positive cells (%) ± SD is shown (b) (
In order to first demonstrate the differential expression patterns of cultured versus resident cells of the cornea, the expression of markers for stemness, mesenchymal, epithelial, endothelial cell origin, and extracellular matrix components and adhesion proteins was carried out in the native cornea (Figure
Immunofluorescent staining of normal human full thickness corneal sections. Images were acquired at 10x (left column) and 40x (middle and right columns) magnifications for each marker. Proteins and markers were stained by Alexa Fluor 488 conjugated secondary antibodies (green). DAPI (blue) counterstaining was applied to visualize the cell nuclei.
Distribution of corneal stroma markers in the various corneal regions.
Marker | Peripheral stroma | Central stroma | Anterior stroma | Posterior stroma |
---|---|---|---|---|
ABCG2 | − | − | − | − |
ABCG5 | − | − | − | − |
ALDH1A1 | ++ | ++ | ++ | + |
|
++ | ++ | ++ | + |
CD31 | + | + | + | + |
CD34 | ++ | ++ | ++ | ++ |
CD73 | − | − | − | − |
CD90 | − | − | − | − |
CD105 | − | − | − | − |
Collagen I | ++ | ++ | ++ | ++ |
Collagen IV | − | − | − | − |
CXCR4 | − | − | − | − |
Fibroblast marker | − | − | − | − |
Fibronectin | − | − | − | − |
Ki-67 | − | − | − | − |
Nestin | − | − | − | − |
VE-Cadherin | − | − | − | − |
Vimentin | ++ | ++ | ++ | ++ |
“−” stands for no staining, “+” for a medium intensity signal, and “++” for a strong staining.
Expression of ABCG2 and ABCG5 could not be detected in situ in any of the corneal layers or regions. Strong and similar staining for ALDH1A1 and
The expression of the major stromal component, Collagen I, showed strong positivity, while absence of Collagen IV could be detected throughout the stroma. The triad of MSC markers: CD73, CD90, and CD105 was negative throughout the native cornea, and markers like Ki-67, CXCR4, and Nestin could not be detected in the tissue either. The presence of Vimentin could be confirmed in the stroma, while Fibronectin appeared to be negative in the native corneal stroma. The expression of an antireticulocyte, fibroblast marker, and VE-cadherin was also found negative in this tissue (Figure
FACS analyses revealed a high expression of MSC markers: CD73, CD90, and CD105, with no significant difference among the culture conditions (Figure
Percent of positive cells for given surface markers in the four different conditions ± SD is shown (a) (
CD73, CD90, and CD105/Endoglin were significantly expressed higher in the cultured CSCs compared to the native stroma (14-, 95-, and 25-fold;
Gene expression profile of the native corneal stroma compared to cultured CSCs. Relative quantities are shown on a logarithmic scale, in the native and pooled cultured cells from the different isolation/cultivation methods; mean relative quantity (RQ) ± standard deviation (SD) is represented. Significance values are depicted as
ABCG2 was expressed 60-fold more in primary CSCs compared to the native stromal cells (
Cells derived from the corneal stroma can be a good source for corneal research, drug testing, and future cell therapy purposes in the eye or other organs [
The shortage of donor corneas worldwide [
Immunostaining of the native corneal tissue revealed no significant difference in the expression of previously described markers in the central versus peripheral parts of the cornea. The native corneal stroma showed no expression of the putative stem cell marker—the efflux protein ABCG2—while a strong staining was observed in the cultures from both the central and peripheral regions produced by both techniques of isolation (explant versus enzymatic). This difference in the expression found in the cultured CSCs versus the native cells could also be confirmed at the gene expression level. Upregulation of ABCG2 may result in a stronger resistance of cultured cells to externally applied therapy (e.g., chemotherapy), while cancerous cells have been known to exploit use of such molecules to survive harsh conditions [
Expression of ALDH1A1, a corneal crystalline, is essential for the maintenance of transparency, downregulation of which is associated with corneal haze [
All layers of the cornea expressed
CD31 is usually expressed by vascular endothelial cells and is likely to be involved in leukocyte migration. The role of CD31 in the attraction and adhesion of polymorphonuclear cells in corneal wound healing has been demonstrated before [
Little is known about the pleiotropic functions of CD34, which is often referred to as the marker of hematopoietic progenitors, while evidence suggests this marker to likely function in immunological processes, such as regulating migration and mobility of eosinophil granulocytes and dendritic cells, as demonstrated in animal knockout experiments [
The triad of MSC markers—CD73, CD90, and CD105—was found to be negative in situ, in contrast to the strong positivity observed in all cultured cells ex vivo. The same results were found when comparing the gene expression of the cultured versus native cells. These findings are supported by recent findings in another independent study [
Characterizing the extracellular matrix which makes the backbone of the cornea is also very important to elucidate the difference between the corneal stroma cells
These findings further strengthen the cultured cells respond to a change in the environmental niche surrounding them, which is likely compensated by deposition of de novo synthesized collagen ex vivo (data not shown).
Fibronectin was also not present in the native, intact corneas, which confirms the corneal wound healing properties of this extracellular matrix component. Deposits of Fibronectin have been shown to appear in the epithelium and stroma soon after penetrating trauma, although it disappears over the course of two weeks [
The presence of Ki-67 could not be detected in the native corneal sections either. This further confirms there was no trauma affecting the control, native epithelium, or the other layers of the cornea, thus indicating presence of induced cell proliferation, although both explant and enzymatic technique generated cultures from the different corneal regions displayed actively proliferating Ki-67-positive cells.
The mesenchymal marker Vimentin was found to be expressed in the stroma. This is in line with the findings of others. Vimentin has also been shown in knockdown studies to cause development of corneal haze [
KLF4, an important stemness marker, has been associated with tissues exposed to the outside world [
Altogether, a general upregulation of stemness and mesenchymal cell markers was observed in the ex vivo cultivated CSCs, with a downregulation of function-related molecules, which should all be considered when treating such cells as MSCs. The potential of corneal stromal cells has been demonstrated numerous times in animal studies before; however, it seems like certain compromises should be made when using such cells as part of
The present study shows no phenotypic or genotypic difference between CSCs produced by the digestion or explant methods from the central or peripheral regions of the cornea. However, the gene expression and protein profile of native corneal stroma cells compared to ex vivo expanded CSCs shows that the latter likely adapt from an
The authors declare that they have no conflicts of interest.
Richárd Nagymihály and Zoltán Veréb contributed equally to this work.
Richárd Nagymihály has been supported by an EU COST Action BM1302 travel grant to perform parts of their experiments at the host site at the Norwegian Center for Stem Cell Research/Center for Eye Research, University of Oslo, Norway. Goran Petrovski and the Stem Cells and Eye Research Laboratory, Department of Ophthalmology, Faculty of Medicine, University of Szeged, Hungary, have been supported by the National Brain Research Program (KTIA_NAP_13-A_III/9 and GINOP-2.3.2-15-2016-00006) (Hungary), cofinanced by the EU and the European Regional Development Fund. Furthermore, the project was also funded through the Center for Eye Research (CER), Department of Ophthalmology, Oslo University Hospital and University of Oslo and the Norwegian Financial Mechanism 2009–2014 under Project Contract no. MSMT-28477/2014, Project 7F14156.