Toxicological evaluation of preservative-containing and preservative-free topical prostaglandin analogues on a three-dimensional-reconstituted corneal epithelium system

Abstract

Aims: Using an established three-dimensional (3D) toxicological model based on reconstituted human corneal epithelium (HCE), this study investigated the tolerability of four topical intraocular-pressure-lowering agents: the commercial solutions of benzalkonium chloride (BAC)-containing 0.005% latanoprost, 0.004% travoprost, 0.03% bimatoprost containing 0.02%, 0.015% and 0.005% BAC, respectively, and the preservative-free (PF) tafluprost. Solutions of 0.01% and 0.02% BAC alone were also evaluated for comparison.

Methods: The 3D-HCEs were treated with solutions for 24 h followed or not by a 24 h recovery period. We used a modified MTT (3(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) procedure to assess cell viability in the HCE. Frozen sections of HCE were analysed using fluorescence microscopy for the evaluation of apoptosis (terminal deoxynucleotidyl transferase mediated dUTP nick end labelling), inflammation (ICAM-1) and proliferation (Ki67). Corneal epithelial tight junctions (occludin and tight junction protein 1 (zona occludens 1)) were also assessed by en face confocal microscopy in response to the different eye-drops.

Results: The MTT test revealed that the cytotoxicity of antiglaucoma eye-drops was primarily related to the concentration of their common BAC preservative (0.02% BAC-latanoprost>0.015% BAC-travoprost>0.005% BAC-bimatoprost). PF-tafluprost did not induce any obvious cytotoxicity, showed the least expression of inflammatory or apoptotic markers and revealed preservation of membrane immunostaining of tight junction proteins in comparison with BAC-containing solutions.

Conclusion: The toxicological model of the 3D reconstructed corneal epithelia model confirmed the ocular surface cytotoxicity of BAC-containing antiglaucomatous solutions. Compared with the formulations containing the toxic preservative BAC, PF-tafluprost was well tolerated without inducing significant corneal epithelium deterioration.

Figure 1

Cell-viability MTT 3(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) test: cell viability of phosphate-buffered saline (PBS)-, preservative-free (PF)-tafluprost-, 0.005% benzalkonium chloride (BAC)-bimatoprost-, 0.010% BAC-, 0.015% BAC-travoprost-, 0.020% BAC-latanoprost- and 0.020% BAC-treated 3D-human corneal epithelium after 24 h incubation without or with a 24 h recovery period. Antiglaucoma treatments induced a BAC-dose-dependent decrease in cell viability as assessed by the MTT test. PF-tafluprost showed a higher level of viability than all BAC-containing eye-drops. *p<0.01 compared with PBS at the same time point; #p<0.03 compared with 0.010% BAC at the same time point; $p<0.002 or p<0.03 ($$) compared with 0.020% BAC at the same time point; ◆p<0.05 compared with 0.020% BAC-latanoprost at the same time point; ●p<0.02 compared with 0.015% BAC-travoprost at the same time point.

Figure 2

Apoptosis analysis: localisation of terminal deoxynucleotidyl transferase mediated dUTP nick end labelling (TUNEL)-positive cells (green) on 3D-human corneal epithelium samples after phosphate-buffered saline (PBS) (A), preservative-free (PF)-tafluprost (B), 0.005% benzalkonium chloride (BAC)-bimatoprost (C), 0.010% BAC (D), 0.015% BAC-travoprost (E), 0.020% BAC-latanoprost (F) and 0.020% BAC (G) incubation after 24 h without (left column) or with the 24 h recovery period (right column). Quantification of TUNEL-positive cells (H) was performed after 24 h of treatment with all solutions followed or not by a 24 h recovery period. At 24 h or 24 h+24 h recovery, no or very few apoptotic cells were observed after PBS (A) or PF-tafluprost (B) treatments. Bimatoprost with 0.005% BAC (C) induced moderate expression of apoptosis, and 0.010% BAC (D), 0.015% BAC-travoprost (E), 0.020% BAC-latanoprost (F) 0.020% BAC (G) all induced a great number of TUNEL-positive cells principally in the apical cell layers, and also in the middle epithelial layers. The nuclei were stained with 4′,6-diamidino-2-phenylindole (blue). The quantification of TUNEL-positive cells shows the increase in apoptotic cell number in a BAC dose-dependent manner. BAC at 0.010% and 0.020% significantly increased the number of TUNEL-positive cells at 24 h, and at 24 h+24 h recovery compared with control. The antiglaucoma eye-drops travoprost and latanoprost, which contain higher concentrations of BAC, showed a much higher expression of apoptotic cells than did PF-tafluprost or control. *p<0.02 for 0.010% BAC and p<0.01 for other solutions compared with PBS at the same time point; #p<0.005 compared with 0.010% BAC at the same time point; $p<0.001 compared with 0.020% BAC at the same time point; ◆p<0.003 compared with 0.020% BAC-latanoprost at the same time point; ●p<0.03 compared with 0.015% BAC-travoprost at the same time point.

Figure 3

Inflammation analysis: immunolocalisation of ICAM-1 (CD54) positive cells (green) on 3D-human corneal epithelium samples after phosphate-buffered saline (PBS) (A), preservative-free (PF)-tafluprost (B), 0.005% benzalkonium chloride (BAC)-bimatoprost (C), 0.010% BAC (D), 0.015% BAC-travoprost (E), 0.02% BAC-latanoprost (F) or 0.020% BAC (G) treatments after 24 h of treatment without or with the 24 h postincubation period. Quantification of ICAM-1 (H) was performed after 24 h of treatment with all solutions followed or not by a 24 h postincubation period. ICAM-1-expressing cells were already found in the PBS (A) and PF-tafluprost (B) groups with the same level of expression after 24 h or 24 h+24 h. The increase in ICAM-1 expressions was observed after the treatments with 0.005% BAC-bimatoprost (C) and 0.010% BAC (D) with and without the 24 h postincubation period. However, few remaining ICAM-1 positive cells were observed after treatments with 0.015% BAC-travoprost (E), 0.020% BAC-latanoprost (F) and 0.020% BAC (G) on the 3D-human corneal epithelium samples owing to the erosion. The nuclei were stained with propidium iodide (red). The quantification of terminal deoxynucleotidyl transferase mediated dUTP nick end labelling-positive cells confirmed the observations. *p<0.001 compared with PBS at the same time point; #p<0.005 compared with 0.010% BAC at the same time point; $p<0.0001 compared with 0.020% BAC at the same time point; ◆p<0.002 compared with 0.020% BAC-latanoprost at the same time point; ●p<0.008 compared with 0.015% BAC-travoprost at the same time point.

Figure 4

Proliferation analysis: immunolocalisation of Ki-67 positive cells (green) on 3D-human corneal epithelium samples after phosphate-buffered saline (PBS) (A), preservative-free (PF)-tafluprost (B), 0.005% benzalkonium chloride (BAC)-bimatoprost (C), 0.01% BAC (D), 0.015% BAC-travoprost (E), 0.020% BAC-latanoprost (F) or 0.020% BAC (G) treatments after 24 h of treatment without (left column) or with the 24 h recovery period (right column). Ki67 immunostaining showed an epithelium with proliferating cells not restricted to the basal layer after PBS treatment (A) at 24 h. The same aspect was found after the treatment of PF-tafluprost (B). Numerous proliferating cells, with a great number located in all the layers, were found after the treatments of 0.005% BAC-bimatoprost (C), 0.010% BAC (D), 0.015% BAC-travoprost (E), 0.020% BAC-latanoprost (F) and 0.020% BAC (G) at 24 h. At 24 h+24 h recovery, we observed the persistence of some proliferative cells after the treatments of PBS or PF-tafluprost. However, at this time, there were no more proliferative cells after the treatments with all BAC-containing solutions.

Figure 5

Tight junction: occludin and tight junction protein 1 (zona occludens 1, ZO-1). Immunofluorescence analysis of occludin (green: A to G) and ZO-1 (green: H to N) expressions using en face confocal microscopy after treatment with phosphate-buffered saline (PBS) (A), preservative-free (PF)-tafluprost (B), 0.005% benzalkonium chloride (BAC)-bimatoprost (C), 0.010% BAC (D), 0.015% BAC-travoprost (E), 0.020% BAC-latanoprost (F) and 0.020% BAC (G) treatments after 24 h (left column) or 24 h+24 h recovery (right column). The nuclei were stained with 4′,6-diamidino-2-phenylindole (blue). After the treatment with PBS or PF-tafluprost, the 3D-human corneal epithelium revealed a thin occludin/ZO-1 immunostaining in the most superficial cells of samples by forming a green ring around, leaving a diffuse cytoplasmic staining. We still found a slight tight junction expression in the apical cells with 0.005% BAC-bimatoprost (C). The expressions disappeared completely after the treatments with 0.010% BAC (D), 0.015% BAC-travoprost (E), 0.020% BAC-latanoprost (F) and 0.020% BAC (G) at 24 h or 24 h+24 h recovery, leaving dense green patches. Note that the nuclei became smaller after treatments with high-dose BAC-containing solutions.