Differentiation of rabbit bone marrow mesenchymal stem cells into corneal epithelial cells in vivo and ex vivo

Abstract

Purpose: To examine whether bone marrow mesenchymal stem cells (MSCs) could be differentiated into corneal epithelial cells in vivo and ex vivo.

Methods: In vivo, BrdU labeled rabbit MSCs (Rb-MSCs) were suspended in the fibrin gels and transplanted onto the surface of the damaged rabbit corneas. Histology and molecular phenotype were studied on postoperative day 28. In vitro, labeled Rb-MSCs were cultured for three days in two different systems: (1) Group A: Rb-MSCs were co-cultured with rabbit limbal stem cells (Rb-LSCs) by the Transwell culture system. A suspension of Rb-LSCs was added to the upper membrane surface, and the inserts were positioned in the culture wells, which were incubated with Rb-MSCs; (2) Group B: Supernatant medium that had first been used to culture Rb-LSCs and then filtered with a 0.45 mum filter was used to culture Rb-MSCs. For both groups, immunofluorescence and flow cytometric analysis were used to examine the expression of cytokeratin 3 (CK3) in differentiated Rb-MSCs.

Results: In vivo, the data showed that following transplantation of Rb-MSCs, the rabbit's damaged corneal surface was successfully reconstructed and that some Rb-MSCs participated in the healing of the injured corneal epithelium and expressed CK3. In vitro, the data showed that Rb-MSCs rapidly differentiated into cells with a morphological and molecular phenotype of corneal epithelial-like cells. For both groups, the differentiated Rb-MSCs were positive for corneal epithelial-specific marker CK3. In Group A, flow cytometry analysis showed that at day one, only 3.46+/-1.9% of cells expressed CK3. This increased to 7.24+/-3.80% at day two and decreased slightly (5.50+/-3.33%) at day three. The proportion of CK3 in Group B was 4.09+/-1.84% at day one, rising to 9.31+/-5.92% after 24 h, but falling (4.37+/-2.61%) at day three. The mean differences are significant between each group and the negative control, but was not significant between Group A and Group B.

Conclusions: MSCs could differentiate into corneal epithelial-like cells in vivo and ex vivo.

Figure 1

Characterization of Rb-LSCs and Rb-MSCs in vitro. A light photomicrograph of cultured Rb-LSCs is seen in panel A (200X). Positive staining of p63 (B) and integrinβ1(C) were observed on Rb-LSCs (200X). CK3 was positive on few of Rb-LSCs (D) and Hoechst 33342 (E) was used as a counterstain (200X). A light photomicrograph of cultured Rb-MSCs is seen in panel F (200×). Under the adipogenic induction medium cultured for 14 days, Rb-MSCs showed a positive reaction with oil red O stain (G) (200X). H: Immunofluorescent staining showing Rb-MSCs labeled by BrdU expressed green fluorescence in the cell nucleus (200X).

Figure 2

Flow cytometric analysis of the Rb-MSCs. The data showed that 97.5% of cells were positive for CD29 (A) and 5.15% of cells were positive for CD34 (B). C and D: Negative controls.

Figure 3

Characterization of the fibrin gel and the rabbit corneas after transplantation. Hematoxylin and eosin staining of Group 1 (A) showed the fibrin gel with Rb-MSCs and Group 2 (B) the fibrin gel without cells (400X). A slit-lamp photograph (C) and hematoxylin and eosin staining (D) showed the model of limbal stem cell deficiency (400X). A slit-lamp photograph of the rabbit corneas after transplantation showed the opacification and neovascularization in Group 1 (E) and in Group 2 (F). Hematoxylin and eosin staining showed that goblet cells, new vessels, and inflammatory cells were present in some regions in Group 1 (G) and Group 2 (H; 400X). Positive CK3 staining was continuous throughout whole corneal epithelium in Group 1 (I), but was irregular in Group 2 (J) by immunofluorescent staining (100X). Double staining showed corneal epithelial cells expressed BrdU (green) and CK3 (red) in Group 1 (K; 200X). A higher magnification of the double staining is seen in panel L (400X).

Figure 4

Morphological and phenotypic of differentiated Rb-MSCs in vitro. Double staining showed expression of CK3 and BrdU both in Group A (A-F) and Group B (H-M) as of day three. Only BrdU positive staining showed in negative controls (G,N). The red staining was CK3 and the green color was BrdU (400X).

Figure 5

Flow cytometry analysis of the ratio of CK3+ cells in all groups. Data represent the mean±SE% of results in three replicate experiments (The asterisk indicated a p<0.05, when compared with control data). There were no significant differences between Group A and Group B (p>0.05).