Hysterographic and histological changes during the normal oestrous cycle of the domestic cat have been well described (Chatdarong et al. 2005 ; Zambelli and Cunto 2005). Reported oestrogen-induced changes in the reproductive tract, both at the macroscopic and microscopic level, appear relatively consistent between mammalian species, especially when dose-, tissue- and time-dependent variation is accounted for. The type of change, if any, elicited following isoflavone administration may provide some indication of their mechanism of action and their potential oestrogenic or anti-oestrogenic capacity may be evaluated from histological changes associated with their administration.
1 .7.4.2.a. The Vagina
Changes reported in the domestic cat vagma during the follicular phase (involving elevated circulating E2 concentrations) have included a reduction of cervix height, relaxation of the cervix, increased dorsal medial folding, and elongation of the vaginal vault (fornix) (Arthur et al. 1 989; Zambelli and Cunto 2005). During the luteal phase (when P4 is elevated), the vaginal mucosa receives decreased blood supply, while secretions are also seen to decrease (Arthur et al. 1 989; Feldman and Nelson 1 996). The vaginal epithelium has been shown to become cornified, and mucus to be secreted in greater quantities under the influence of oestrogen, compared to during quiescent states (Arthur et al. 1 989; Feldman and Nelson 1 996).
1 .7.4.2.b. The Uterus
Studies have reported that, whilst under the influence of E2 during the follicular phase, the feline uterus and uterine horns becomes enlarged with a congested, hyperaemic endometrium, and has increased secretory activity in preparation for implantation (Arthur et al. 1 989; Feldman and Nelson 1 996; Chatdarong et al. 2005) . During the oestrous cycle, the endometrium was at its thickest in the follicular phase, although no significant difference was found between luteal and follicular phase thickness (Chatdarong et al. 2005). Myometrial thickness was significantly greater in both the luteal and follicular phase, compared to the inactive phase (Chatdarong et al. 2005) .
Luminal epithelial cell height was significantly greater during follicular and luteal phases, compared to the inactive phase, but glandular epithelium cell height was increased most significantly in the luteal phase (Chatdarong et al. 2005). Corresponding structural changes occurred in the luminal epithelium, which became pseudostratified in the follicular phase, and progressed to being hyperplastic in the luteal phase (Chatdarong et al. 2005). The glandular epithelium was pseudostratified in the luteal phase, with only minimal changes observed in the follicular phase (Chatdarong et al. 2005). The degree of glandular proliferation, secretion and dilatation was also greater in the luteal phase, and
cystic formation was slightly elevated (Arthur et al. 1 989; Feldman and Nelson 1 996; Chatdarong et al. 2005).
However, despite the apparent differentiation and proliferation in the reproductive tract during oestrus, cellular proliferation (as indicated by proliferative cell nuclear antigen) was unaffected by oestrous stage (Chatdarong et al. 2005). The high degree of variation observed in this parameter within stages of the cycle was the most likely reason for the lack of significant differences.
1 .7.4.2.c. The Oviducts and Ovaries
Minimal follicular growth was evident during anoestrus or inter-oestrus in feline tissue, and any corpora lutea present were regressed and non-functional (Arthur et al. 1 989; Feldman and Nelson 1 996). However, during oestrus, ovaries become enlarged with raised vesicular follicles visible on their surface (Arthur et al. 1 989; Feldman and Nelson 1 996). Oviducts become hypertrophied and fully ciliated with increased secretion during the fol licular phase of the cycle (West et al. 1 977). Similar effects were observed in OVX cats treated with exogenous E2 (West et al. 1 977), and the mode of basal body and cilia formation in domestic cats appears to be essentially the same as in oviducts of other species (Verhage and Brenner 1 975).
Corpora lutea are reported to be present on the ovanan surface following ovulation (Arthur et al. 1 989; Feldman and Nelson 1 996) and these vesicles release P4. Endogenous and exogenous P 4 has been shown to cause deciliation, decreased cell height, and secretion, in the oviducts of cats (West et al. 1 977). Indication of P4 influence in the pregnant and E2-primed OVX cat treated with exogenous P 4 was inferred by the presence of dense secretory granules, surrounded by light halos in the glandular epithelium of cat uteri (West et al. 1 977).
1 . 7 .
4
.3
.F olliculogenesis
Folliculogenesis in the domestic cat is easily studied due to the presence of large numbers of ovarian follicles at all stages of development in the feline ovary (Bristol-Gould and Woodruff 2006). It is now known that primary oocyte formation in the cat occurs between days 40 - 50 of gestational life, and is completed approximately 8 days after birth (Bristol-Gould and Woodruff 2006). Each primordial follicle contains an oocyte arrested in the first meiotic prophase, which does not resume development until triggered by gonadotrophins (Bristol-Gould and Woodruff 2006).
The TGF-P superfamily has been shown to play an important role in balancing the growth and apoptosis involved in follicular development in the domestic cat (Bristol and Woodruff 2004). lsoflavones are reported to be capable of modulating TGFP expression and activity (see Section 1 .5 .2 .6.b) and this may represent a mechanism for isoflavone activity in the reproductive tract of cats.
Conversely, unlike findings in bovine pregnancy (Woclawek-Potocka et al. 2005ab ), prostaglandin F2a (PGF2a) was not luteolytic during the early luteal phase in the cat (Wildt et al. 1 979). Therefore, if the ability of isoflavones to increase PGF2a seen in cattle (Woclawek-Potocka et al. 2005ab) was also to occur in domestic cats, it is unlikely that this will be translated into ovarian changes.
1 . 7 .
4
.4
.Steroid receptor expression and growth factors
The expression and distribution of steroid receptors has been found to vary according to the stage of oestrous cycle in the domestic cat (West et al. 1 977; Li et al. 1 992). Overall ERa expression in the uterus and oviducts was reported to be low during anoestrus, increased during oestrus, and reduced during pregnancy and the luteal phase (West et al. 1 977). Oestrus cats exhibited significant concentrations of ERa and PR in the luminal and glandular epithelium, as well as stromal fibroblasts (Li et al. 1 992).
In addition to changes in receptor expressiOn, distributional effects have also been observed according to stage of cycle. The luminal surface epithelium was shown to express significant concentrations of ERa or PR only during oestrus (the follicular phase) (Li et al. 1 992). These receptors were not detected in these cell types following ovulation or exogenous P 4 administration (Li et al. 1 992). However, in the glandular epithelium, ERa was expressed in the nuclei of all cells, as well as in the stromal fibroblast cells during both fol licular and luteal phases (Li et al. 1 992), albeit in decreased concentrations during pregnancy (West et al. 1 977). Progesterone receptor expression disappeared from the luminal epithelial cell nuclei fol lowing ovulation, but was maintained in the stromal fibroblast cells (Li et al. 1 992). The decreased ERa concentrations observed during pregnancy were thought to be the result of increasing P 4 concentrations, since E2 concentrations were high enough to maintain ERa (West et al. 1 977). These studies investigating ER expression of the domestic cat uterus were conducted prior to the discovery of ER� (first cloned in the rat by Kuiper et al. 1 996) and therefore no data exists to date as to the distribution and expression of this isoform.
The surface (luminal) epithelium of OVX or pregnant domestic cat uteri is not known to contain ER or PR (Li et al. 1 992). However, this tissue stil l becomes convoluted in response to both endogenous and exogenous E2 during these phases. The glandular epithelium was also responsive to E2 and became hypertrophic and increasingly differentiated (Chatdarong et al. 2005). It has been suggested that the waviness and coiling of the uterine lumen during these phases was a prostagenic effect, related to the proliferation of endometrial glands (Chatdarong et al. 2005). The presence of E Rs and PRs in glandular tissue may explain the pathway for such histological changes. However, since luminal epithelium contains significant concentrations of EGF receptor it is thought that EGF p lays a more important role in modifying the luminal epithelium (which is devoid of ERa and PR), via the interaction of TGF� and EGF with ovarian steroids (Boomsma et al. 1 997). Receptors for EGF are also found in the glandular epithelium. These are l ikely to be significant factors in the steroid-induced changes observed in the
domestic cat reproductive tract during normal oestrous cycles, and following exposure to exogenous hormones ( Boomsma et al. 1 997).
Interestingly, although EGF functions under E2-influence to control the growth and differentiation of domestic cat uteri and oviducts, neither growth factor nor the EGF receptor were affected in expression or distribution by steroid treatment in OVX cats (Boomsma et al. 1 997). The only apparent response to steroid treatment reported in OVX cats was the appearance of dense deposits of TGFB in the epithelium and stroma following P 4 treatment, which has also been noted in other species (Boomsma et al.
1 997). The ability of isoflavones to modulate growth factor activity and expression in other species may therefore not elicit similar changes in the expression of these factors in the cat.
Structural and functional similarities between human and feline ERa supports the use of the cat as a model for human ERa (Cardazzo et al. 2005) . The GenBank (2006) database on homology has reported that feline and human ERa have 95% homology, whilst the human and feline PR also share 95% homology. The homology of feline and mouse PGR was lower (89%), whilst fel ine and mouse ERa share 1 00% homology (GenBank 2006). This is important when extrapolating effects on steroid receptor expression in the cat from rodent or human studies. However, as discussed previously, genomic similarities may not necessarily manifest parallel responses at the tissue or whole-animal level.