Differential Expression of Sex-Steroid Receptors in the Choroid Aligns With Central Serous Chorioretinopathy Sex Prevalence Across Different Ages

Abstract

Purpose: The purpose of this study was to investigate the presence of sex-steroid receptors in human choroidal tissue across different ages and sex, aiming to better understand the pronounced sex difference in central serous chorioretinopathy (CSC) occurrence.

Methods: Paraffin-embedded enucleated eyes of 14 premenopausal women, 15 postmenopausal women, 10 young men (<45 years), and 10 older men (>60 years) were used. A clinically certified immunostaining was performed to detect the presence of the androgen receptor (AR), progesterone receptor (PR; isoform A and B), and estrogen receptor (ERα). The stained slides were scored in a blinded manner for positive endothelial cells and stromal cells in consecutive sections of the same choroidal region.

Results: Our analysis revealed the presence of AR, PR, and ERα in endothelial cells and stromal cells of choroidal tissue. The mean proportion of AR-positive endothelial cells was higher in young men (46% ± 0.15) compared to aged-matched women (29% ± 0.12; P < 0.05, 95% confidence interval [CI]). Premenopausal women showed markedly lower mean proportion of ERα (5% ± 0.02) and PR-positive endothelial cells (2% ± 0.01) compared to postmenopausal women (15% ± 0.07 and 19% ± 0.13; both P < 0.05, 95% CI), young men (13% ± 0.04 and 21% ± 0.10; both P < 0.05, 95% CI), and older men (18% ± 0.09 and 27% ± 0.14; both P < 0.05, 95% CI). Mean PR-positive stromal cells were also less present in premenopausal women (12% ± 0.07) than in other groups.

Conclusions: The number of sex-steroid receptors in the choroidal tissue differs between men and women across different ages, which aligns with the prevalence patterns of CSC in men and postmenopausal women.

Central serous chorioretinopathy (CSC) is a chorioretinal disease associated with high male prevalence,^1–3 having corticosteroid exposure as the most pronounced extrinsic factor,^4,5 and characterized by macular subretinal fluid leakage.⁶ CSC occurs up to six times more frequently in men compared to women.^1–3,7 The mean age at onset of CSC in men peaks in a range of 30 to 54 years of age.^1–3 In women, CSC presents typically at a slightly older age, generally after the menopause, between 50 and 69 years old, and often shows complications, such as secondary choroidal neovascularization.^7,8 Interestingly, pregnancy-associated CSC has been reported, most frequently in the third trimester, which coincides with elevated serum progesterone, estrogen, and cortisol levels.^9,10 Dysfunction of the choroidal vasculature has been thought to be part of the underlying pathology in CSC, based on clinical manifestations diagnosed with multimodal imaging (e.g. indocyanine green angiography and optical coherence tomography to visualize choroidal vasculature).^6,11 The clinical manifestations in eyes with CSC are delayed choroidal filling, choroidal vascular hyperpermeability, pigment epithelial detachments, and accumulation of subretinal fluid. Theories pointing toward the abnormalities of vortex vein outflow, choroidal venous overload, and the effect of a thickened sclera have been proposed to contribute to the pathogenesis of CSC.^11–13 In addition, there is an evidence that the scleral thickness in patients with steroid-associated CSC was thinner than non-steroid-associated cases.¹⁴ However, how corticosteroids may affect choroidal vasculature in the pathogenesis of CSC still remains to be elucidated. Half-dose photodynamic therapy has been proven to be effective in treating CSC, in contrast to micropulse laser and oral mineralocorticoid receptor antagonists.^15–17 The clinical symptoms of CSC mostly resolve spontaneously, however, the visual function remains poor in some cases.¹⁸ 

There might be a role for sex-steroid hormones (androgens, estrogens [E2], and progesterone) in the pathogenesis of CSC, based on the male-female prevalence, and the fact that pregnancy-associated CSC normally resolves after childbirth.⁹ A recent study by Brinks et al. (2022) found a high free testosterone/estradiol ratio in female patients with CSC compared to healthy women.¹⁹ However, an in vitro study using cultured choroidal endothelial cells (ECs) from male and female donors not exposed to hormones did not show intrinsic differences in the cortisol response between the sexes.²⁰ Whether the signaling of sex-steroid hormones in the choroid is involved in the pathogenesis of CSC in men and women is unknown. 

The presence of sex-steroid receptors defines the cellular targets of sex-steroid hormones. The transcriptional effects of androgens, progesterone, and estrogens are mediated by the androgen receptor (AR), progesterone receptor (PR), and estrogen receptor (ER), respectively. To date, the knowledge on the presence of sex-steroid receptors in the human choroid is limited to transcriptional studies that have generally been performed in older donors.^21–25 Importantly, the expression of sex-steroid receptors and their subtypes can change with age in response to hormone fluctuations throughout life, addressing the importance of studying receptor expression in younger donors.²⁶ 

Choroidal ECs have often been implicated in the pathogenesis of CSC. The choroidal stroma, which is defined as the connective tissue to support the stability of Sattler's and Haller's layers and the suprachoroidal layer,²⁷ is less studied in the clinical manifestation of CSC. Yet, a reduction of choroidal stroma area was observed in patients with CSC.²⁸ 

This study aims to screen the immunolocalization of the sex-steroid receptors in the ECs and stromal cells of the choroidal tissue from premenopausal women, postmenopausal women, young men (<45 years), and older men (<60 years) with a certified staining method for clinical diagnosis in oncology. We separately quantified the numbers of immunoreactive ECs and stromal cells and calculated the ratio of ECs to stromal cells for each receptor. 

Materials and Methods

Samples

The samples of enucleated eyes for this study were obtained from the Department of Pathology of the Leiden University Medical Center (LUMC). The details of age, sex, and reason for enucleation can be found in Table 1. Whether CSC had occurred in the medical history of the donors was not known. All samples were obtained from the database collection between the years 2007 and 2022. Eyes with a history of intraocular tumors or cases post-irradiation were excluded, which was assessed by an ophthalmic pathologist (author R.V.). Any history of medication (including the use of contraceptives), menstrual cycle status, and pregnancy history of the subjects were unknown. The study was approved by the LUMC ethics committee (registration number RP23.001) and performed in accordance with the tenets of the Declaration of Helsinki. 

Table 1.

 

Demographics of the Samples

 

Immunohistochemistry

Immunohistochemistry was performed on formalin-fixed paraffin-embedded samples detecting AR, ERα, and PR (detecting both A and B isoforms). All procedures from cutting, slide preparation, providing tissues for staining control, and immunostaining were performed by Erasmus University Medical Center Pathology Research and Trial Service (PARTS; ISO15189:2012 certification) to obtain a standardized result. Immunohistochemistry was performed using a VENTANA automated slide stainer (Roche Tissue Diagnostic, Almere, The Netherlands) according to the protocol from the manufacturer. All antibodies (Roche Tissue Diagnostic)²⁹ were used for clinical diagnosis and listed in Table 2. For scoring purposes, all slides were subsequently scanned by using 3DHistech PANNORAMIC 250 Flash III DX (3DHistech, Budapest, Hungary) with 40× optical magnification at the Light and Electron Microscopy Facility of the LUMC. 

Table 2.

 

List of Antibodies for the Detection of Sex-Steroid Receptors

 

Scoring

The first author (S.G.) scored all samples for the presence or absence of immunoreactivity in the nuclei of respective tissues. The confirmation of cell histomorphology and positive immunoreactivity were discussed and all slides were re-evaluated again in a blinded manner with an ophthalmic pathologist (author R.V.). Both scoring processes by the authors, S.G. and R.V., were performed in a blinded manner to each other's scoring and sex and age of status of the samples. In cases of discordance, a consensus was achieved between the two observers. Positive controls were present on each slide: AR – human testis; ERα and PR – human cervix and normal mammary gland. Scoring was carried out by calculating the percentage of positive-stained nuclei from a minimum of 200 nuclei of each cell type (ECs and stromal cells) per section, irrespective of the staining intensity, using a manual annotation in QuPath (version 0.4.3; Belfast, Northern Ireland, UK)³⁰ and CaseViewer (version 2.4; 3DHistech, Budapest, Hungary) for Windows. Within the choroid, we identified stromal cells and ECs based on their histomorphology in all choroidal layers (choriocapillaris and Haller's and Sattler's layer). Quantification of stromal cells was only performed on non-pigmented cells. Next, the positive ratio of ECs to stromal cells was calculated for each receptor staining, in accordance with the common quantification technique of tumor-stroma ratio in estimating tumor proportion in the histological section. 

Statistical Analysis

The data normality distribution was assessed by using the normality tests method D'Agostino – Pearson. The age difference between groups was tested with an unpaired t-test. The comparison of the percentage of positive-stained nuclei from AR, ERα, and PR across the groups and the calculated ratio ECs/stromal cells were analyzed using 1-way ANOVA, followed by Tukey's multiple comparison tests. Correlation between ECs and stromal cells in individual donors was assessed per group by simple linear regression with the output of regression coefficient (R²) and P. All analyses were performed using GraphPad Prism (version 9.3.1 [471]; Boston, MA, USA). A P < 0.05 with 95% confidence interval (CI) was considered statistically significant. The outputs of significance were shown as * P < 0.01; ** P < 0.002; *** P < 0.0002; and **** P < 0.0001. 

Results

The demographic data of the samples that were included in this study can be found in Table 1. The average of the young women's age was 29.1 ± 6.4 years, whereas this was 70.3 ± 6.4 years for older women. The median age range of natural menopause in women is 50 to 52 years, although it varies in different races and ethnicities,³¹ thus, it is safe to assume that our subjects were pre- and postmenopausal, respectively. An unpaired t-test showed that there were no statistical differences in age between premenopausal women and young men, and postmenopausal women and old men (P = 0.77 and 0.93, respectively). The differences in age were only observed within the female groups (P < 0.0001) and within the male groups (P < 0.0001). 

The positive staining controls in the section of the human testis, mammary gland, and cervix showed an appropriate immunoreactivity for all staining (AR, ERα, and PR, respectively). The cell nuclei were stained in blue, and the nuclei with positive reactivity to AR, ERα, and PR antigen were visualized with the red chromogen (Figs. 1A–C). In the choroid, immunoreactivity for AR, ERα, and PR (Figs. 2A–C and Supplementary Figs. S1–S3, respectively) was observed in ECs and stromal cells, which resided in all choroidal layers. Two slides, from the postmenopausal women and old men groups, contained less than 200 nuclei and were therefore excluded from the calculated percentage ratio of nuclear positivity. 

Figure 1.

 

Nuclear immunoreactivity of sex-steroid receptors in human control tissue. Nuclear immunoreactivity of the androgen receptor was observed in human testis (A); the estrogen receptor alpha was observed in the human mammary gland (B); and the progesterone receptor was observed in the human cervix (C). The negative-stained nuclei are in blue, positive-stained nuclei are in red. Magnification 150×. AR, androgen receptor; ERα, estrogen receptor alpha; PR, progesterone receptor.

Figure 2.

 

The immunoreactivity of sex-steroid receptors in endothelial cells and stromal cells in human choroidal tissue. Nuclear staining for the androgen receptor (A); estrogen receptor alpha (B); and progesterone receptor (C) with magnification 40× (right images) and 150× (left images). Positive-stained nuclei are in red and negative-stained nuclei are in blue. Stromal cells that stained positive are indicated by black arrows, and negative-stained stromal cells are indicated by orange arrows. Endothelial cells that stained positive are indicated by green arrows, and negative-stained endothelial cells are indicated by gray arrows. Higher magnification (300×) for stromal cells, positive-stained (a) and negative-stained (b); for endothelial cells, positive-stained (c) and negative-stained (d) were placed in black, orange, green, and grey squares, respectively. AR, androgen receptor; BM, Bruch's membrane; CC, choriocapillaris; ERα, estrogen receptor alpha; HL/SL, Haller's and Sattler's layers; PR, progesterone receptor; RPE, retinal pigment epithelium.

A higher mean percentage (46% ± 0.15, mean ± standard deviation) of positive nuclei for AR staining was observed in ECs in young men, compared to premenopausal women (29% ± 0.12; * P < 0.01; see Fig. 3A and Supplementary Table S1). There was no statistical difference between the number of AR-positive stromal cells (Fig. 3B). Premenopausal women showed a lower calculated ratio of ECs/stromal positive cells for AR compared to young men (0.67 ± 0.19 and 1.01 ± 0.30, respectively, * P < 0.01; Fig. 3C). In the premenopausal group, AR-positive ECs were correlated with AR-positive stromal cells (R² = 0.5; P < 0.02; Fig. 3D). No correlation between ECs and stromal cells was observed in other groups. 

Figure 3.

 

The quantitation of androgen receptor (AR), estrogen receptor alpha (ERα), and progesterone receptor (PR)-positive nuclei in human choroidal vasculature. The mean level of AR-positive endothelial cells (ECs) (A) and AR-positive stromal cells (B). These values were calculated the ratio of ECs/stromal (C), which showed a comparable result within same sex, but statistical significance was observed between premenopausal women and young age-matched men. A significant correlation between AR-positive ECs and stromal cells from premenopausal women (D). The mean level of ERα-positive ECs (E) and ERα-positive stromal cells (F). The calculated ratio of ERα-positive ECs/stromal (G). The mean level of PR-positive ECs (H) and PR-positive stromal cells (I), which were used for calculated ratio ECs/stromal (J). A strong correlation was observed between the number of PR-positive ECs and PR-positive stromal cells in men (K). Premenopausal women had the lowest ratio of ECs/stromal in both ERα and PR-stained nuclei, meanwhile postmenopausal women and all age-matched men groups showed a comparable result in ERα and PR-stained nuclei. For each bar graph, data represent mean ± standard deviation. Statistical significance was determined by P < 0.05, 1-way ANOVA test, and 95% CI. The outputs of significance were shown as * P < 0.01; ** P < 0.002;*** P < 0.0002; and **** P < 0.0001.

Premenopausal women showed a lower mean percentage level of ERα-positive ECs (5% ± 0.02) compared to mean percentage of postmenopausal women (15% ± 0.07; ** P < 0.002) and all the men groups (13% ± 0.04; * P < 0.01 and 18% ± 0.09; **** P < 0.0001; see Fig. 3E and Supplementary Table S2). Similarly, a lower level of the calculated ratio of ECs/stromal positive cells was demonstrated in the premenopausal group (0.20 ± 0.09), compared to other groups (postmenopausal women = 0.49 ± 0.23, * P < 0.01; young men = 0.46 ± 0.16, * P < 0.01; and older men = 0.67 ± 0.41, *** P < 0.0002; Fig. 3G). ERα-positive ECs were not correlated with ERα-positive stromal cells in any of the groups. 

The mean percentage level of PR-positive ECs (2% ± 0.01) and stromal cells (12 ± 0.07) in premenopausal women was lower than mean percentage of postmenopausal women (ECs = 19% ± 0.13, *** P < 0.0002; stromal cells 24% ± 0.11, * P < 0.01), young men (ECs = 21% ± 0.10, **** P < 0.0001; stromal cells = 25% ± 0.12, * P < 0.01), and older men (ECs = 27% ± 0.14, **** P < 0.0001; and stromal cells = 30% ± 0.11, *** P < 0.0002; see Figs. 3H, 3I, and Supplementary Table S3). The calculated ratio of the ECs/stromal cells for PR-positivity was lower in premenopausal women (0.18 ± 0.16) compared to other groups (**** P < 0.0001; Fig. 3J). A correlation of PR-positive ECs and stromal cells was observed in both young and older men groups, but not in the pre- or postmenopausal women (R² = 0.7; P < 0.0001; Fig. 3K). 

Discussion

In the current study, we used clinically certified staining to describe the presence of sex-steroid receptors in the human choroid tissue comparing men and women across different ages in order to find a possible explanation for sex differences in CSC occurrence. Overall, our data indicate differences in sex-steroid receptors expression between the sexes and age groups in choroidal ECs but not in stromal cells. Premenopausal women had a substantially lower immunoreactivity of ERα and PR in the ECs compared to all other groups, and the count of AR-positive ECs was higher in young men compared to aged-matched women. Therefore, ECs are the candidate cell type in which differential expression of sex hormones may contribute to CSC disease risk. 

The higher prevalence of CSC in men compared to women argues for androgens as a risk factor. Circulating testosterone may act via the AR, and potentially via ERs (ERα or ERβ) after intracellular conversion to estradiol if the enzyme aromatase is also expressed.³² The data from the younger subjects suggest that higher expression of AR and of ERα may confer risk to develop CSC for younger men, compared to premenopausal women. The prevalence of CSC in women follows different patterns than in men, as CSC manifests more in older women and occasionally during pregnancy. Our finding of higher PR expression in postmenopausal women compared to premenopausal women suggest that this may be a risk to CSC in women. In addition, the abundance of PR increases in the last trimester of pregnancy, as studied in the myometrium tissue.³³ Progesterone levels in blood reach a peak between age 10 years and 40 years in women and are strikingly high during pregnancy.³⁴ Thus, it may well be possible that progesterone plays a role in pregnancy-associated CSC. 

The association between the expression of sex-steroid receptors and CSC development is not straightforward to interpret. Free sex-steroid hormones in men and women can influence the cellular level of sex-steroid receptors, which is commonly known as homologous regulation.³⁵ Despite the fact that the menstrual cycle of all women subjects was not documented in this study, we may assume that exposure to estrogens and progesterone was higher in the premenopausal group, compared to the postmenopausal samples. Therefore, the low number of ERα/PR-positive cells may indicate homologous downregulation as a consequence of strong ERα/PR-mediated signaling in the ECs. Indeed, short-term contraceptive use decreases the level of ERα and PR in breast parenchyma of premenopausal women.³⁶ However, the literature seems to argue against this mechanism in the human vasculature.^37–39 In one study, the abundance of the ERα protein in peripheral antecubital veins did not elevate in estrogen-deficient postmenopausal women,³⁹ in contrast to our present data. If homologous downregulation is indeed involved, it occurred in a cell-type specific manner,^40,41 given that the reduction of ERα and PR-positive cells occurred only in ECs and not stromal cells in premenopausal women. 

AR was earlier shown to induce its own expression, including in human ECs, which illustrates a homologous upregulation of AR.^42,43 The level of free and total testosterone decreases and sex hormone binding globulin (SHBG) increases across the lifespan of men.⁴⁴ In contrast, the total testosterone level slightly changes throughout the women's lifetime,⁴⁵ but other androgens, such as dehydroepiandrosterone (DHEA) and androstenedione, decrease steeply in aging women.^45,46 This predicts that the AR would decrease in older men, which may be relevant to our findings. Interestingly, compared to young men, premenopausal women had a substantially lower number of AR-positive ECs. This group also had a positive correlation with AR-positive stromal cells. Of note, one study showed that functional response to androgen exposure in relation to angiogenesis is sex-dependent.⁴⁷ This sex-dependence to androgen exposure disappeared upon overexpression of AR in the female derived ECs. Therefore, even if CSC development is not directly related to androgen levels,^19,48 unresponsiveness to any circulating androgens in young women could potentially explain that CSC occurs more in young men. 

Our data showed that premenopausal women had a very low average number of PR-positive ECs and stromal cells compared to postmenopausal women. In the literature, there is few data on the regulation of PR by progesterone in ECs. Our data indicates that any PR signaling in choroidal tissue would not be mediated by ECs in premenopausal women. In another tissue, the activation of PR by progesterone in vascular endothelium has been associated with barrier instability and vascular permeability.⁴⁹ Interestingly, PR-positive ECs and stromal cells were positively correlated in both male groups, indicating that there could be a potential interaction between PR-positive ECs and stromal cells involved in the choroidal vascular leakage in CSC. These findings suggest that choroidal PR may be involved in the signaling of choroidal vasculature. 

In light of these findings, the absence of ERα and PR signaling in the choroidal tissue from premenopausal women may be protective. However, in case the low expression of choroidal ERα/PR does reflect homologous downregulation (even if there is a lack of supporting literature), then ERα/PR signaling in choroidal ECs may actually be protective for women. Another interpretation of these findings is that if there is a lack of ERα/PR signaling in premenopausal choroidal ECs, then possible activating protective effects of E2 or progesterone would not be directly mediated via choroidal ECs. Regardless, the striking differences that we observed suggest that sex hormones and sex-steroid receptors should be further included in the model to study the pathogenesis of CSC. 

Our study has some limitations. Due to privacy regulations, it was impossible to obtain a health status of human donors that exactly reveals sex hormone homeostasis (e.g. menstrual cycle, contraceptive use, medication, and obesity). The ECs and stromal cells in the choroidal tissue were identified based on their histomorphological features, which is relatively common practice in cancer research. However, when there is no abnormality, there is a risk of misclassification with other cell types, for example, pericytes and smooth muscle cells. To distinguish each cell type, double staining with the correct cell marker would have been necessary, which we did not perform due to limited material. Next, the digitalized slides’ scoring was performed regardless of the intensity of the staining and so our analysis hinges on the absence/presence of the receptor expression, and did not take the expression level of the receptors into account. Last, we did not determine ERβ expression because ERβ antibody is not routinely used in the clinical pathology. 

In conclusion, our study emphasizes that sex and age are associated to the density of sex-steroid receptors in the human choroid tissue. Young men may be more vulnerable to CSC as a relatively high number of sex-steroid receptors respond to male sex hormones. In contrast, young women may confer protection through choroidal ECs that are less responsive to female sex hormones. To unravel pathophysiologic mechanisms behind the striking difference in prevalence of CSC between men and women, further in vitro studies may focus on the effect of sex hormones (androgen or progesterone) in induced pluripotent stem cells derived ECs from patients with CSC - providing additional information on how sex hormones change EC functionality. 

Acknowledgments

The authors would like to thank Mieke Versluis, PhD, and the biobank archive of the Department of Pathology (LUMC) for the sample database and sample collection. 

Supported by the following foundations: Stichting Macula Fonds, Retina Nederland Onderzoek Fonds, Stichting BlindenPenning, Algemene Nederlandse Vereniging ter Voorkoming van Blindheid, and Landelijke Stichting voor Blinden en Slechtzienden, which contributed through UitZicht, as well as Rotterdamse Stichting Blindenbelangen, Stichting Leids Oogheelkundig Ondersteuningsfonds, Haagse Stichting Blindenhulp, Stichting Ooglijders, ZonMw VENI Grant, and Gisela Thier Fellowship of Leiden University (CJFB). The funding organizations had no role in the design or conduct of this research. They provided unrestricted grants. 

Disclosure: S. Galuh, None; O.C. Meijer, None; J. Brinks, None; R.O. Schlingemann, None; C.J.F. Boon, None; R.M. Verdijk, None; E.H.C. van Dijk, None