Selective effects of estradiol on human corneal endothelial cells

Abstract

Fuchs endothelial corneal dystrophy (FECD) results from genetic and environmental factors triggering mitochondrial and oxidative stress in corneal endothelial cells (CEnCs) leading to CEnC death and corneal opacification. FECD is more common in women than men, but the basis for this observation is unknown. Because FECD is commonly diagnosed around the time of the menopausal transition in women when estrogen levels decrease precipitously, we studied the effects of the potent estrogen,17-β estradiol (E2) on growth, oxidative stress, and metabolism in primary cultures of human CEnCs (HCEnCs) under conditions of physiologic 2.5% O ₂ ([O ₂ ] _2.5 ) and under hyperoxic stress ([O ₂ ] _A : room air + 5% CO ₂ ). We hypothesized that E2 would counter the stresses of the hyperoxic environment in HCEnCs. HCEnCs were treated ± 10 nM E2 for 7-10 days at [O ₂ ] _2.5 and [O ₂ ] _A followed by measurements of cell density, viability, reactive oxygen species (ROS), mitochondrial morphology, oxidative DNA damage, ATP levels, mitochondrial respiration (O ₂ consumption rate [OCR]), and glycolysis (extracellular acidification rate [ECAR]). There were no significant changes in HCEnC density, viability, ROS levels, oxidative DNA damage, OCR, and ECAR in response to E2 under either O ₂ condition. We found that E2 disrupted mitochondrial morphology in HCEnCs from female donors but not male donors at the [O ₂ ] _A condition. ATP levels were significantly higher at [O ₂ ] _2.5 compared to [O ₂ ] _A in HCEnCs from female donors only, but were not affected by E2. Our findings demonstrate the overall resilience of primary HCEnCs against hyperoxic stress. The selective detrimental effects of hyperoxia and estradiol on HCEnCs from female but not male donors suggests mechanisms of toxicity based upon cell-sex in addition to hormonal environment.

Fig 1.. Estrogen receptor gene expression in human corneal endothelium by qRT-PCR.

mRNA expression for ERα, ERβ, and GPER, (and housekeeping/control genes AQP1, GPDH, CANX, LDHA, RPLP0, and UBC) in non-diseased human corneal endothelium (n= 4, corneas #4–7), and FECD patient corneal endothelium (n=3, corneas #1–3). Data presented are the mean of three reverse transcription reactions ± standard deviation. Donor numbers correspond to those in Table 1.

Fig 2.. Estrogen receptor protein expression in human corneal endothelium.

Western blot for: (A) GPER and (B) ERβ in native human corneal endothelium (n=1 donor for GPER, cornea #8; n=2 donors for ERβ, corneas #9, 10), and PC3 (n=3 for each antibody), and MCF7 (n=3 for each antibody) cells lines. Theoretical molecular weight of GPER is 42 kDa and ERβ is 55–60 kDa. (C) En face immunofluorescence localization of ERβ in normal human corneal endothelium (representative images from corneas #11–13). Donor numbers correspond to those in Table 1.

Fig 3.. Effects of G1 and E2 on PC3 and HCEnC on cell growth.

(A) Number of PC3 cells with 1 nM, 10 nM, and 100 nM E2 or 0.1 μM, 1.0 μM, 10 μM G1 treatment (n=12). P-values from t-test compared to the control: * p<0.01, ** p<0.001. (B) HCEnC counts with 10 nM E2 or 1.0 μM G1 treatment (n=5, corneas #14–17). Data presented as mean ± SD of total cell counts per well. P-values from ANOVA single factor.

Fig 4.. Cell viability assay of HCEnC in the presence and absence of oxygen stress, and E2 and G1 treatment.

(A) Cell viability measurements (log₂ luminescence/cell) with E2 or G1 treatment in HCEnCs (n=5, corneas #14–17). Data is presented as mean ± SD. *p-value from ANOVA. Each plot in B-G represent the data for a single donor with p-value from paired t-tests. Donor numbers correspond to those in Table 1.

Fig 5.. Representative images of mitochondrial morphology grading.

Mitochondrial arbor from MitoTracker Red stain was graded for each cell as diffuse (normal), intermediate, or fragmented (abnormal). Nuclei were stained with DAPI (blue). Representative images from corneas #27R-34L.

Fig 6.. Mitochondrial morphology in HCEnCs in the presence and absence of oxygen stress, and estradiol treatment.

(A) Immunofluorescence images of mitochondrial arbor, stained with MitoTracker Red, in HCEnCs following H₂O₂ treatment. (B-C) Quantification of mitochondrial arbor grade as diffuse (normal), intermediate, or fragmented (abnormal) following H₂O₂ treatment in HCEnCs (n=4, corneas #27R, 27L, 28L, 33R). P-value from t-test compared to the control. (B) Data presented as mean ± SD. (C) Donor numbers correspond to those in Table 1. (D) Quantification of mitochondrial morphology grade with the presence or absence of 10 nM estradiol treatment at [O₂]_2.5 or [O₂]_A. P-value from ANOVA single factor. Pairwise comparisons were performed by Tukey post hoc analysis, with significant differences highlighted in red. N= 9 total, 4 males, and 5 females (corneas # 27R-33L).

Fig 7.. Oxidative damage in HCEnCs in the presence and absence of oxygen stress, and estradiol treatment.

(A) Percentage of 8-oxo-dG positive HCEnCs in the presence or absence of estradiol treatment (10 nM) at [O₂]_2.5 or [O₂]_A. Data presented as mean ± SD. P-value from ANOVA single factor. N= 8 total, 4 males, and 4 females (corneas #27L-33L). (B) Immunofluorescence imaging of 8-oxo-dG in PC3 and HCEnCs following H₂O₂ treatment. (C) Percentage of 8-oxo-dG positive cells following H₂O₂ treatment in HCEnCs (n=4, corneas #27R, 27L, 28L, 33R). Donor numbers correspond to those in Table 1.

Fig 8.. OCR and ECAR in the presence and absence of oxygen stress, and E2 treatment.

(A) OCR and (B) ECAR measurements in the presence or absence of E2 treatment (10 nM) at [O₂]_2.5 or [O₂]_A. (C-J) Trends in mean data values with and without E2 treatment at both [O₂]_2.5 and [O₂]_A, by cell sex for each condition. X-axis for time is a total duration of 41 minutes for OCR and 52 minutes for ECAR recordings. Data presented as mean ± SD. N= 5–6 total, 2–3 males, and 2–3 females (corneas #40R-45).