Cell culture
Human SKPs were successfully isolated and presented as small floating spheres. CEC-like cells were well achieved using our previous differentiation method. The CEC-like cells displayed hexagonal or polygonal shapes and formed a mosaic monolayer, which were similar to HCEC. These cells strongly expressed HCEC major markers, Na+/K+ ATPase and ZO-1(Fig.1 A, B).
Long-term observation of CEC-like cells transplantation into the monkey corneal endothelial dysfunction model
In the experimental group, the cornea remained transparent, and almost no keratic precipitates or corneal neovascularization appeared (Fig.2 A). In addition, the pupil was round with a positive light reflex, the iris texture was clear, the anterior chamber depth was normal, the aqueous fluid was clear, and the lens was transparent. In the control group, the cornea remained opaque with edema and as a consequence the iris could not be seen (Fig.2 B). Clinical observation of all five monkeys are shown in Supplemental figure 2.
The corneal thickness in the experimental group remained steady within the range of 0.5 mm - 0.55 mm from 1 month to 2 years after transplantation. The corneal thickness in the control group continued to be above 1.0 mm 2 years after transplantation, and neovascularization appeared (Fig.3 A, C). The cells after transplantation were in the form of a polygonal monolayer (Fig.3 B). The total endothelial cell density continued to increase in the first three months, peaking at an average of 3022 cells/mm2, 84% of pre-operation. But after three months it gradually fell. The cell density rapidly decreased in the first year and slowly in the second year. It decreased by about 20% during the first year and by 5% during the second year compared to the third month. However, the cell density was still more than 2000 cells/mm2 within 2 years, 63% of pre-operation (Fig.3 D). However, because of corneal opacity, this could not be detected in the control group. Other ophthalmic examinations including intraocular pressure, chamber angle, B-ultrasonography, and fundus imaging showed no apparent pathological abnormity (Fig.3 E, F, G).
Polymerase chain reaction and DNA-sequencing after transplantation
PCR was carried out to show the existence of human or monkey DNA in the samples. We designed primers for human and monkey specific Cytb gene fragments to identify human or monkey DNA. With the human Cytb gene fragments primer, a 459 bp gene fragment in HCEC group was obtained (Fig.4A HCEC). With the monkey Cytb gene fragments primer, a 130 bp gene fragment in MON group was obtained (Fig.4A MON). But, with the two primers, only a 459 bp but not a 130 bp gene fragment in CEC-like cells was obtained (Fig.4A CEC-like). However, with the two primers, both a 459 bp and a 130 bp gene fragment in the EXP1 group were obtained (Fig.4A EXP1), the same as the EXP2 group (Fig.4A EXP2).
DNA-sequencing of the 459 bp gene fragment products in EXP2 group was carried out. It showed that the base pairs sequence highly corresponded to that of the designed human Cytb DNA fragment (Fig.4 B). This indicated that the amplified DNA products acquired in the experimental group had a high homology with human DNA, which proved the long-term existence of injected CEC-like cells in the 1-year and 2-year experimental groups (EXP1 and EXP2).
Histological examination after transplantation
Immunofluorescent staining of frozen section showed Na+/K+ ATPase expression in EXP2 group, indicating the persistent pump function (Fig.4 C). Through HE staining it could be seen that the corneal thickness in the EXP2 group was similar to the normal group. Under magnified observation, the CEC-like cells tightly adhered to the posterior surface of the cornea in a monolayer. The morphology of the corneal epithelium, stromal cells, and endothelium in the EXP2 group were all similar to the normal group (Fig.4 D, E). The EXP1 data was same as EXP2, and the EXP1 data are not shown. However, in the control group, the cornea was very thick and edematous. Under magnified observation, the Descemet’s membranes were bare and almost no cells were detected. The corneal stromal layer had severe edema and a lot of inflammatory cells. Collagen fibers were irregularly arranged, layered, and broken. Stromal cells had an abnormal morphology. The corneal epithelial layer was also vacuolate (Fig.4 F).
Functional pattern of gene expression in HCEC, CEC-like cells, EXP1 cells, and EXP2 cells
Venn analysis was carried out to show co-expressed and specially expressed genes between samples or groups. The GO annotation analysis showed the top 20 GO sets of abundance in specially expressed genes, including classification of the biological process, cellular component, and molecular function.
The Venn analysis of HCEC and CEC-like cells showed that 71.6% of all expressed genes were co-expressed. The number of specific genes of CEC-like cells and HCEC was 1671 and 3186, accounting for 9.8% and 18.6% (Fig.5 A). The GO annotation analysis of specific genes in the two groups showed that the abundance ranking in the two groups was almost the same with only few differences. The percent of the signaling and developmental process ranked higher in CEC-like cells than that of HCEC. The percent of the extracellular region ranked higher than the catalytic activity in CEC-like cells, whereas in HCEC they were the opposite (Fig.5 B).
The Venn analysis of CEC-like/EXP1/EXP2 cells showed that 79.3% of all expressed genes were co-expressed. The number of specific genes of CEC-like/EXP1/EXP2 cells was 356/508/519, accounting for 2.1%/3.0%/3.1%, respectively (Fig.5 C). The GO annotation analysis of specific genes in the three groups showed that the abundance ranking was similar with only a few differences. The percent of signaling and positive regulation among biological process ranked higher in CEC-like cells. However, the macromolecular complex and extracellular region among cellular component ranked higher in EXP1 and EXP2 (Fig.5 D).
The Venn analysis of HCEC/EXP1/EXP2 cells showed that 68.4% of all expressed genes were co-expressed. The number of specific genes of HCEC/EXP1/ EXP2 cells was 2119/392/414, accounting for 19.1%/2.1%/2.2%, respectively (Fig.5 E). Through GO annotation analysis it was found that the abundance ranking was similar to that in CEC-like/EXP1/EXP2 cells GO annotation analysis (Fig.5 F).
The correlation analysis was carried out to detect the relativity between samples or groups. The r value between each biological repeat sample of HCEC and CEC-like cells was more than 0.98 (data not shown), demonstrating good biological repeatability. The mean r value between CEC-like cells and EXP1 or EXP2 cells was both about 0.954. The r value between EXP1 and EXP2 cells was 0.991. The mean r value between CEC-like cells and HCEC was about 0.87, but after 1-2 years transplantation the r value between EXP1 or EXP2 cells and HCEC was 0.873 and 0.874, respectively (Fig.5 G).
Differential expression gene analysis of CEC-like cells, EXP1 cells, and EXP2 cells
Expression variance analysis was used to detect differential expression genes (DEGs). Genes with FDR < 0.05 and |log2FC| ≥ 1 were considered as significant. GO enrichment analysis was carried out to obtain the GO Term of significant DEGs enrichment.
4566 significant DEGs were detected between CEC-like cells and EXP1 cells, among which 3385 (74.1%) genes were up-regulated and 1181 (25.9%) were down-regulated. The GO enrichment analysis showed that the significant DEGs in EXP1 cells vs CEC-like cells mainly took part in regulation of water loss, mitotic recombination, sensory organ morphogenesis, and kidney development (Fig.6 B C).
4524 significant DEGs were detected between CEC-like cells and EXP2 cells, among which 3360 (74.3%) genes were up-regulated and 1164 (25.7%) were down-regulated. Significant DEGs in EXP2 cells vs CEC-like cells mainly took part in cell-cell junction organization/assembly, cell adhesion, chromosome segregation, pattern specification, and DNA replication (Fig.6 D E).
19 significant DEGs were detected between EXP1 cells and EXP2 cells, among which 12 genes were up-regulated and 7 were down-regulated. The genes could not be GO enriched.