Fig 1.
RA signaling pathway regulates the dynamic balance of differentiation and proliferation in LECs.
Retinoic acid (RA) is the downstream metabolite of retinol, which is initially taken up into the cell by STRA6 (1). Once inside, retinol binds to cellular retinol-binding protein 1 (CRBP1), initiating a two-step oxidative process that converts it into functional RA (2, 3). Excess retinol can follow three alternative pathways: it can be stored in the cytoplasm as retinyl ester (8), secreted paracrinally to adjacent tissues (7), or degraded via CYP26A1 (4). As carrier proteins, cellular RA-binding protein 2 (CRABP2) and fatty acid-binding protein (FABP5) maintain a dynamic balance, both binding RA to transport it into the nucleus for downstream gene transcription [15,16]. RA transported by CRABP2 binds to the RXR-RAR heterodimer at the PPRE region, regulating apoptosis and suppressing cell growth (5) [15,16]. Alternatively, RA can activate the FABP5-bound PPARG axis, inducing transcription at the RARE region, which is essential for maintaining basic cell functions and promoting cell growth (6) [18,21,22]. STRA6: stimulated by retinoic acid 6, SDR: short-chain dehydrogenase/reductase, LRAT: lecithin retinol acyltransferase, PPRE: peroxisome proliferator-activated receptor response element, RARE: retinoic acid response element.
Table 1.
Demographic information of corneal donors.
Table 2.
Primer pairs used for qPCR.
Table 3.
Antibodies used for western blot analysis.
Fig 2.
PAX6 mRNA and protein levels (b,c) in control siRNA (blue bars) and in PAX6 knockdown (red bars) limbal epithelial cells (LECs) without and after 1µM or 5µM RA incubation for 48 hours.
mRNA values are shown as geometric mean±geometric standard deviation (SD). Total protein staining was used for protein normalization of each lane. The total protein normalization factor TPN is indicated below each lane. Band intensity from WB is shown with mean±SD. Two-way ANOVA, followed by Dunnett’s test was used. Significant p values (<0.05) are highlighted. PAX6 mRNA (a) and protein levels (b) were significantly lower after PAX6 siRNA knockdown than using control siRNA (p < 0.001 for both). Nevertheless, there was no significant difference between any other groups without or using different RA concentrations (p ≥ 0.611).
Fig 3.
mRNA and protein expression of enzymes in the retinoic acid signaling pathway in control siRNA (blue bars) and in PAX6 knockdown (red bars) limbal epithelial cells (LECs) without and after 1µM or 5µM RA incubation for 48 hours.
RDH10 (a), ADH7 (d), ALDH1A1 (g), CYP26A1 (j) mRNA and RDH10 (b), ADH7 (e), ALDH1A1 (h) protein levels (Western blot) are shown. A representative western blot of RDH10 (c), ADH7 (f), and ALDH1A1 (g) is displayed. mRNA results are shown as geometric mean ± geometric standard deviation (SD), and protein results are shown as mean ± standard deviation, with significant differences indicated. Two-way ANOVA, followed by Dunnett`s test has been used. Including all analyzed groups, ADH7 and ALDH1A1 mRNA expression levels (p = 0.018, p = 0.015) and protein levels (p = 0.003, p < 0.001) were significantly downregulated in PAX6 knockdown LECs, compared to control siRNA treated LECs. ALDH1A1 mRNA levels were significantly downregulated after 5 µM RA treatment in the control group (p = 0.038). In addition, after 1 µM and 5 µM RA treatment, CYP26A1 mRNA expression was upregulated in control (p < 0.001, p < 0.001) and PAX6 knockdown LECs (p = 0.001, p = 0.002). Nevertheless, there was no significant difference between any other groups without or using different RA concentrations (p ≥ 0.077).
Fig 4.
mRNA and protein expression of binding proteins in control siRNA (blue bars) and in PAX6 knockdown (red bars) limbal epithelial cells (LECs) without and after 1µM or 5µM RA incubation for 48 hours.
CRABP2 (a), FABP5 (d), RBP1 (g) mRNA and CRABP2 (b), FABP5 (e) protein levels (Western blot) are shown. A representative western blot of CRABP2 (c) and FABP5 (f) is displayed. mRNA results are shown as geometric mean ± geometric standard deviation (SD), and protein results are shown as mean±standard deviation, with significant differences indicated. Two-way ANOVA, followed by Dunnett`s test has been used. Including all analyzed groups, mRNA and protein expression levels of FABP5 were significantly lower in PAX6 knockdown LECs, than in control siRNA-treated LECs (p < 0.001, p < 0.001). Following 5 µM RA treatment, CRABP2 mRNA level in the PAX6 knockdown group (p = 0.02) and CRABP2 protein level in both the control and the PAX6 knockdown groups was downregualted (p = 0.003, p = 0.02). Nevertheless, there was no significant difference between any other groups without or using different RA concentrations (p ≥ 0.067).
Fig 5.
mRNA and protein expression of receptors in RA signaling pathway in control siRNA (blue bars) and in PAX6 knockdown (red bars) limbal epithelial cells (LECs) without and after 1µM or 5µM RA incubation for 48 hours.
PPARG (a), RXRA (d), RXRB (g), RARA (h), RARB (i) mRNA and PPARG2 (b), RXRA (e) and RARB (j) protein levels (Western blot) are shown. A representative western blot of PPARG2 (c), RXRA (f), and RARB (k) is displayed. mRNA results are shown as geometric mean ± geometric standard deviation (SD), and protein results are shown as mean ± standard deviation, with significant differences indicated. Two-way ANOVA, followed by Dunnett`s test has been used. PPARG2 mRNA level was significantly upregulated and PPARG2 and RARB protein levels were significantly downregulated in PAX6 knockdown LECs, compared to control siRNA-treated LECs (p = 0.012, p = 0.007, p < 0.001). Following 5µM RA treatment, in the control group, RARB mRNA expression and RXRA protein expression were downregulated (p = 0.007, p = 0.007). After 1µM and 5µM RA treatment, RARB protein level was downregulated in the control group (p = 0.02, p = 0.004). After 5µM RA treatment, in the PAX6 knockdown group, RARA and RARB mRNA levels were significantly downregulated (p = 0.023, p < 0.001). Nevertheless, there was no significant difference between any other groups without or using different RA concentrations (p ≥ 0.054).
Fig 6.
mRNA and protein expression of VEGFA in control siRNA (blue bars) and in PAX6 knockdown (red bars) limbal epithelial cells (LECs) without and after 1µM or 5µM RA incubation for 48 hours.
VEGFA mRNA (a) and protein (Western blot) (b) expression are shown. mRNA results are shown as geometric mean ± geometric standard deviation (SD), and protein results are shown as mean ± standard deviation, with significant differences indicated. Two-way ANOVA, followed by Dunnett`s test has been used. Following 5µM RA treatment, VEGFA mRNA expression levels were significantly upregulated in controls (p = 0.041). Nevertheless, there was no significant difference between any other groups without or using different RA concentrations (p ≥ 0.127).
Fig 7.
BrdU assay in control siRNA (blue bars) and in PAX6 knockdown (red bars) limbal epithelial cells (LECs) without and after 1µM or 5µM RA incubation for 48 hours.
Two-way ANOVA, followed by Dunnett’s test was used. Significant p values (<0.05) are highlighted. BrdU assay shows a significantly lower proliferation rate in the PAX6 siRNA knockdown group than in control siRNA group (p < 0.001). 5µM RA treatment significantly decreased proliferation rate in controls (p = 0.025). Nevertheless, there was no significant difference between any other groups without or using different retinol concentrations (p ≥ 0.455).
Table 4.
mRNA and protein expression changes in siRNA PAX6 knockdown LECs following retinol and retinoic acid (RA) treatment. Some mRNA and protein level changes were related to the PAX6 knockdown, some others to the treatment, as described in the table.
Table 5.
mRNA expression level changes in conjunctival and corneal epithelial cells following retinol, retinoic acid and RAR antagonist treatment.