Estrogen Protects Lenses against Cataract Induced by Transforming Growth Factor-β (TGFβ)

17-β-estradiol (1,3,5[10]-Estratriene-3, 17β-diol) and progesterone (4-Pregnene-3, 20-dione) were obtained from Sigma Chemical Co. (St. Louis, MO) and human recombinant TGFβ2 from Genzyme Corp. (Cambridge, MA). In some experiments, lenses were derived from normal 6–10-mo-old adult male and female Wistar rats, killed by CO₂ asphyxiation before removal of eyes. Alternatively, ovariectomies were performed on 3-mo-old female Wistar rats under anaesthesia as described elsewhere (14). After waiting for 4–5 wk to ensure clearance of residual estrogen and progesterone, ovariectomized rats were given three daily injections of 0.5 μg 17-β-estradiol or 5 mg progesterone, dissolved in benzyl alcohol/peanut oil (1:3, vol/vol). Control rats received vehicle alone. 1 d later, rats were sacrificed by a lethal dose of nembutal (Boehringer Ingelheim, Sydney, Australia) before removal of eyes.

Lenses were carefully dissected free from surrounding ocular tissues in culture medium, as described previously (10), and cultured with TGFβ2 at final concentrations of 0.025–4 ng/ml, which was added immediately unless otherwise indicated. Controls received no TGFβ. Culture medium was renewed every 2 d throughout the culture period, without further addition of TGFβ. Lenses were cultured for 7 d and the anterior surface was photographed daily to record development of opacities. At the end of the culture period, lenses were fixed in Carnoy's fixative (acetic acid/ethanol, 1:3, vol/vol) and embedded in paraffin.

Generally, serum-free medium 199 containing antibiotics and 0.1% bovine serum albumin, as already described (10), was used as culture medium. When testing the effect of estrogen in vitro, medium 199 was replaced by phenol red-free minimum essential medium (Sigma Chemical Co.).

TGFβ-induced lens opacification begins as diffuse clouding on the anterior surface of the lens. As the response progresses, these regions condense to form distinct opacities leaving a reduced area of clouding. At low concentrations of TGFβ, few distinct opacities are observed at day 7 of culture, and a large proportion of the lens surface remains cloudy. In contrast, most of the initially cloudy areas condense to form numerous distinct opacities at high concentrations (as in Fig. 1 A). On the basis of these observations, a method for measuring the extent of opacification has been developed (15).

Micrographs of lenses at day 7 of culture were used to determine lens opacification. Each micrograph was scanned with an x-ray scanner (3CX; XRS Corp., Torrance, CA) using Adobe Photoshop and XRS Omni Media software. A series of measurements was then made using NIH Image v1.52. In some micrographs, flared reflections of the light source made it impossible to assess the extent of opacification in certain regions (for example see Fig. 1 B). Only micrographs in which the assessable area represented >75% of the total area were used. The assessable area (A) and, within this area, the total area of clouding (B), and the total number of distinct opacities (C), were measured. An “opacification index” was then calculated as follows:

Serial sections of paraffinembedded lenses were processed for routine histology or for immunohistochemical localization of α-smooth muscle actin or type I collagen, as already described (10). Representative lenses from each treatment group were examined by routine histology (all TGFβ2 concentrations) and immunolocalisation (0.15 ng/ml TGFβ2).

Lenses from male rats developed distinct anterior opacities when cultured with 0.15 ng/ml TGFβ2 (Fig. 1,A). In contrast, lenses from female rats remained transparent under these conditions (Fig. 1,B), as did control lenses from male and female rats cultured without TGFβ (not shown; see reference 10). However, at a higher concentration of TGFβ2 (1 ng/ml), lenses from female rats also developed opacities (Table 1). For both sexes, the response increased significantly with concentration of TGFβ.

Histological examination of lenses cultured with 0.15 ng/ml TGFβ2 revealed distinct subcapsular plaques, containing spindle-shaped cells and extracellular matrix (see reference 10) in the lenses from males (not shown). In contrast, the lenses from females and controls retained normal lens architecture, with a monolayer of epithelial cells overlying the fiber mass. Immunolocalization of α-smooth muscle actin and type I collagen showed that, for males, strong reactivity for both of these cataract markers was present in plaques induced by culturing with 0.15 ng/ml TGFβ2, whereas corresponding lenses from females showed no reactivity for α-smooth muscle actin and only very weak reactivity for type I collagen in a few cells in the epithelium (not shown). These markers were not detectable in sections of control lenses from males or females cultured for ⩽7 d without TGFβ (not shown).

Having established that lenses from female rats showed more resistance to the cataractogenic effects of TGFβ than those from males, ovariectomized rats were used to assess the contribution of ovarian hormones to this phenomenon. Lenses from ovariectomized rats (without hormone replacement) developed opacities when cultured with 0.15 ng/ml TGFβ (Table 2; Fig. 2,A), a concentration shown to have negligible effect on lenses from normal female rats (Table 1; Fig. 1,B). Lenses from ovariectomized rats which received estrogen, however, did not develop opacities under these conditions (Table 2; Fig. 2,B), while the response of lenses from rats treated with progesterone was similar to that of lenses from vehicle-treated rats (Table 2; Fig. 2 C).

Histologically, the opacities observed in ovariectomized rats that only received vehicle corresponded with subcapsular plaques or clumps of abnormal cells (Fig. 3,A). Reactivity for the cataract markers α-smooth muscle actin and type I collagen was observed predominantly within the subcapsular plaques (Fig. 3, C and E). In contrast, lenses from estrogen-treated rats retained normal cellular morphology (Fig. 3,B), and no reactivity for α-smooth muscle actin or type I collagen was detected (Fig. 3, D and F). In all these respects, lenses from rats that received progesterone replacement were indistinguishable from lenses from rats that received vehicle alone.

A variety of more subtle histological changes were observed in lenses from ovariectomized rats that did not receive estrogen. Swelling of cortical fiber cells with evidence of degeneration, reminiscent of cortical cataract (11), was commonly observed (for example see Fig. 3,A), generally in the region of the lens anterior to the equator. In addition, nucleated cells were observed migrating along the posterior capsule towards the posterior pole (Fig. 4,A); these cells showed reactivity for type I collagen (Fig. 4,C), but not for α-smooth muscle actin. None of these changes were observed in lenses from estrogen-treated ovariectomized rats, which remained transparent (Fig. 4, B and D). In lenses from normal male and female rats (not shown), concentrations of TGFβ >0.15 ng/ml were required to induce changes as pronounced as those in Figs. 3,A and 4 A. Thus, lenses from normal rats of either sex seemed to be more resistant to the effects of TGFβ than those from ovariectomized rats that did not receive estrogen.

Estrogen administered in vivo as in the above experiments may exert its effects on lens cells directly or indirectly. To determine whether estrogen is capable of influencing lens cells directly, lenses from ovariectomized rats were treated with estrogen in vitro before and during culture with TGFβ (Table 3). Estrogen-treated lenses showed negligible opacification in response to TGFβ. While a small region of haziness was observed at the center of some of these lenses within 6 d of adding TGFβ, no tendency to condense into discrete opacities with time was noted (not shown). Numerous distinct opacities developed in corresponding lenses cultured in parallel without the addition of estrogen, which were thus highly susceptible to the cataractogenic effects of TGFβ under these conditions (Table 3).

The present study grew out of our findings that TGFβ induces lens epithelial explants and cultured lenses to undergo molecular and morphological changes that are typically associated with human subcapsular cataracts (8–10, 16). Both TGFβ1 and TGFβ2 induce cataractous changes, TGFβ2 being the more potent isoform (16). The cataractogenic effects of TGFβ are blocked by a pan-specific antibody against TGFβ (17), and are not simply a “growth factor” effect. Weanling lenses cultured with another growth factor, FGF-2 (or FGF-1), at concentrations known to induce lens cell proliferation and/or fiber differentiation (18) remain transparent (Hales, A., C. Chamberlain, J. McAvoy, unpublished data).

In cultured lenses from weanling rats, opacities induced by TGFβ correspond with subcapsular plaques of aberrant cells including spindle-shaped cells which are often associated with wrinkling of the lens capsule. Abnormal extracellular matrix deposition also occurs, mainly in the plaques. All these changes are typical of anterior and posterior subcapsular cataract and aftercataract (19–22). In addition, TGFβ induces the accumulation of α-smooth muscle actin and type I collagen in both explants and cultured lenses (9, 10). These proteins, which are not generally found in the lens, are present in human anterior subcapsular cataract and in aftercataract (23–26). The present study shows that in lenses from adult rats, as for weanlings, TGFβ induces opacities with morphological and molecular features of cataract. In each case, the plaques associated with the opacities are morphologically indistinguishable from early stage human anterior subcapsular cataracts (11).

Major findings of the present study are that the ovarian hormone, estrogen, protects rat lenses against TGFβ-induced cataract, and that susceptibility to the cataractous changes induced by TGFβ is sex dependent. Culturing lenses from ovariectomized females with TGFβ resulted in marked opacification of the lens. Estrogen replacement in vivo prevented this response, but progesterone replacement did not (Table 2; Fig. 2). Furthermore, lenses from male rats were found to be more susceptible to the cataractogenic effects of TGFβ than those from normal females (Table 1; Fig. 1), and lenses from normal rats of either sex seemed more resistant to the effects of TGFβ than those from ovariectomized rats (Tables 1 and 2). The latter results are also consistent with a protective role for estrogen since circulating estrogens are present in male rats, albeit at much lower levels than in normal females (27), but not in ovariectomized females (28). Further support comes from a recent study involving transgenic mice. Females expressing a mutant, “nonresponsive” estrogen receptor progressively develop cortical and nuclear cataracts (29). It seems unlikely that the greater susceptibility of male rats to the cataractogenic effects of TGFβ is due to the presence of testosterone since, in aging males, cataract increases (5) while testosterone decreases (30).

Sex-dependent and estrogen-related differences in susceptibility to cataract formation, consistent with a protective role for estrogen, have been noted in epidemiological studies. After menopause, the prevalence of cortical cataracts in females suddenly increases relative to males of equivalent age (2, 3). Before menopause, the prevalence of cataract seems to be similar in males and females (3, 31). In addition, the prevalence and severity of certain forms of cataract is lower in postmenopausal women on hormone replacement therapy involving administration of estrogen with or without progesterone, than in those who are not undergoing hormone replacement. This is true whether menopause occurs naturally or as a result of hysterectomy (5–7). Delayed menopause also seems to protect against cataract (5–7). Cernea et al. have suggested a possible link between presenile cataract development and ovarian hormone insufficiency (32). Furthermore, analysis of the age of patients who undergo cataract surgery indicates that men tend to develop presenile cataracts approximately four years earlier than women (2). The present study of cultured rat lenses provides strong support for this epidemiological evidence that estrogen protects against cataract.

This study goes beyond previous studies, however, in that it provides direct evidence that estrogen protects against cataract. Furthermore, it shows that estrogen does so by protecting lens cells against the cataractogenic effects of TGFβ. This protective effect is observed whether estrogen is administered in vivo or in vitro. Thus, it seems likely that estrogen confers protection against cataract by targeting lens cells directly when administered in vivo, although the possibility of additional indirect benefits is not excluded. In other cellular systems, estrogen has been shown to have a variety of effects on TGFβ. It may suppress the effects of TGFβ. For example, estrogen-induced tumorigenesis in the anterior pituitary of rats is accompanied by a loss of sensitivity to TGFβ1 in tumor cells and downregulation of the TGFβ type II receptor (33). Similarly, treating rats with estrogen in vivo reduces TGFβ binding in the uterus, possibly via downregulation of TGFβ receptors (34). On the other hand, estrogen may enhance TGFβ activity when included in vitro (35, 36) or have no effect (37).

The present study adds weight to an increasing body of evidence that hormone replacement therapy not only relieves acute menopausal symptoms, but also offers protection against various diseases (38). These include cardiovascular diseases such as myocardial infarction, atherosclerosis, and related hypertension and stroke, as well as osteoporosis and associated fractures (38–41). Ocular pressure decreases progressively in both women and men, undergoing treatment with estrogen (42). In addition, estrogen use dramatically reduces the risk of idiopathic macular hole formation, an ocular condition that is more common in untreated postmenopausal women than in men of the same age (43). Many postmenopausal women are now receiving estrogen replacement therapy, in conjunction with progesterone where appropriate, but it is not universally advocated or available. In 1994, it was reported that only 5–10% of menopausal women in the United States were receiving this treatment (38).

The opacities that form in rat lenses cultured with TGFβ are anterior subcapsular cataracts. Marked similarities between TGFβ-induced cataract and aftercataract have already been noted (8, 10). Evidence of subtle changes typical of posterior subcapsular cataract and cortical cataract is provided by the present study. We report for the first time that TGFβ can induce migration of aberrant cells along the lens capsule towards the posterior pole (as in Fig. 4,A). A similar abnormal migration of nucleated cells along the posterior capsule is thought to be the basis of human posterior subcapsular cataract formation (20). Some evidence of TGFβ-induced cataract-like change was also observed in the cortical fibers in the present study (Fig. 3 A).

TGFβ is potentially available to lens cells in situ. TGFβ and its mRNA have both been detected in the mammalian eye (44–46), and biologically active TGFβ has been found in samples of the ocular media that bathe the lens obtained from human patients suffering from cataract (47–50). Large reservoirs of inactive TGFβ are present in the ocular media (47, 49, 51, 52) which may become activated, e.g., after eye surgery (53). It is not clear what level of TGFβ stimulation lens cells receive under normal conditions in situ, and little is known about factors that regulate its activity in the eye (17). Estrogens have been detected in the ocular media in untreated female and male rabbits, but it is not clear what levels of active hormone are present (27).

The present study provides further evidence that TGFβ is involved in the aetiology of various forms of subcapsular cataract, and possibly cortical cataract. Our finding that estrogen, administered in vivo or in vitro, protects against cataract is consistent with epidemiological studies. Thus, as suggested by previous studies from this laboratory, the rat lens model appears to be a valid tool for investigating factors that influence cataractogenesis. The parallel between estrogen effects in the rat system and epidemiological evidence also raises the possibility that estrogen may provide protection against human cataract by influencing a TGFβmediated mechanism of cataractogenesis. This study represents another significant step towards elucidating the molecular basis of cataract, the foundation for developing new strategies for prevention or treatment of this debilitating disease.