Impact of Continuous Versus Discontinuous Progesterone on Estradiol Regulation of Neuron Viability and Sprouting After Entorhinal Cortex Lesion in Female Rats

Abstract

Because the estrogen-based hormone therapy (HT) in postmenopausal women typically contains a progestogen component, understanding the interactions between estrogens and progestogens is critical for optimizing the potential neural benefits of HT. An important issue in this regard is the use of continuous vs discontinuous hormone treatments. Although sex steroid hormone levels naturally exhibit cyclic fluctuation, many HT formulations include continuous delivery of hormones. Recent findings from our laboratory and others have shown that coadministration of progesterone (P₄) can either attenuate or augment beneficial actions of 17β-estradiol (E₂) in experimental models depending in part upon the delivery schedule of P₄. In this study, we demonstrate that the P₄ delivery schedule in combined E₂ and P₄ treatments alters degenerative and regenerative outcomes of unilateral entorhinal cortex lesion. We assessed how lesion-induced degeneration of layer II neurons in entorhinal cortex layer and deafferentation in dentate gyrus are affected by ovariectomy and treatments with E₂ alone or in combination with either continuous or discontinuous P₄. Our results demonstrate the combined efficacy of E₂ and P₄ is dependent on the administration regimen. Importantly, the discontinuous-combined E₂+P₄ regimen had the greatest neuroprotective efficacy for both end points. These data extend a growing literature that indicates qualitative differences in the neuroprotective effects of E₂ as a function of cotreatment with continuous versus discontinuous P₄, the understanding of which has important implications for HT in postmenopausal women.

Combined estrogen and progestogen hormone therapies (HTs) are widely used for the treatment of menopause-related climacteric symptoms. In addition to relief from menopausal symptoms, ovarian hormones also have well-established neuroprotective, neurotrophic, and cognitive benefits that may be useful in the prevention of Alzheimer's disease (AD). Despite a wealth of data supporting the independent neuroprotective and neurotrophic effects of estradiol (E₂) and progesterone (P₄), interactions between E₂ and P₄ in combined HT regimens on neural parameters are not well understood. Recent evidence suggests P₄ may modulate the neural actions of E₂ (1). Understanding E₂ and P₄ interactions on neuroprotective outcomes has important implications for combined HT use in postmenopausal women. An important issue in this regard is the effect of continuous vs discontinuous P₄ delivery schedules, which may play an important role in determining the cognitive benefit of a combined HT.

Although sex steroid hormone levels naturally exhibit cyclic fluctuations, many HT formulations include continuous delivery of hormones (2). Prior experimental studies have found that P₄ attenuates the effects of E₂ in some models of neuroprotection (3–8) but increases E₂ neuroprotection in other paradigms (9–11). These apparent discrepancies may reflect divergent effects of continuous vs discontinuous P₄ administration regimens. Specifically, continuous P₄ exposure for prolonged periods is often associated with attenuation of the beneficial E₂ effects (3, 4, 12, 13), whereas discontinuous P₄ administered in combination with E₂ tends to yield improved outcomes (10, 11, 14, 15). Highlighting the importance of the E₂/P₄ delivery regimen, behavioral inhibition or facilitation by concurrent or sequential administration of E₂/P₄ is observed on lordosis (16), cocaine self-administration and sensitization (17), and wheel running (18).

In this study, we examine the effects of P₄ delivery schedule in the presence and absence of E₂ on both degenerative and compensatory changes after the unilateral transection of the perforant path [entorhinal cortex lesion (ECL)] in female rats (Figure 1A). The perforant path is a network of axons running from layer II of the lateral entorhinal cortex (EC) projecting to the molecular layer of the dentate gyrus (DG) of the hippocampal formation. This pathway plays an important role in memory consolidation and is severely affected in AD (19, 20). Selective degeneration of EC neurons and perforant path axons occurs early in AD pathogenesis, accompanied by compensatory axonal sprouting in response to deafferentation of the hippocampus (21). ECL is a model of entorhinal cortex degeneration in AD, with transection of the perforant path resulting in rapid loss of axotomized entorhinal cortex layer II neurons (22) and compensatory synaptic sprouting in the molecular layer of the dentate gyrus (23–25) (Figure 1). Compensatory sprouting is a potentially important regenerative mechanism as it can spare memory deficits associated with hippocampal deafferentation after ECL (25, 26). A significant advantage of the ECL knife lesion model is that it avoids seizures induced by electrolytic lesions or excitotoxic lesions, which would potentially confound interpretation of neuroprotective effects of P₄ due to its anticonvulsant effects. To determine the effects of P₄ delivery schedule on neuroprotection, we compared the effects of E₂ alone or in combination with either continuous or discontinuous P₄ on survival of layer II entorhinal cortex neurons and compensatory neuronal sprouting in dentate gyrus following ECL lesion.

Figure 1.

Unilateral entorhinal cortex lesion model. A, Schematic representation of the entorhinal cortex/perforant path lesion model. Perforant path (PP) axons arise from the distinctly laminated layer II entorhinal cortex neurons (ECII) and terminate on the dendrites of the granule cells in the molecular layer of the DG. Unilateral lesion of the perforant path results in ipsilateral degeneration of axotomized ECII neurons, evidenced by reduced NeuN immunoreactive cells in the ECII region of the lesioned (ipsilateral, B) vs unlesioned hemisphere (contralateral, C). Neurite fiber density in the DG molecular layer, as assessed by Holmes staining, is reduced after the lesion (unlesioned, D; lesioned, E) but shows compensatory regrowth. Compensatory sprouting arises from reinnervation from multiple pathways including commissural/associational afferents, septohippocampal afferents, contralateral EC afferents, and local interneurons (71). Scale bar, 50 μm. hf, hippocampal fissure; ml, molecular layer; gcl, granule cell layer.

Materials and Methods

Animals and surgical procedures

Three-month-old Sprague Dawley rats were purchased ovariectomized (OVX) or sham-OVX from Harlan Laboratories. Seven days after the surgery, rats were administered two consecutive 30-day cycles of hormone treatment (Figure 2). Hormone delivery consisted of continuous E₂, continuous P₄, discontinuous P₄ (10 d/mo), or a combinations of continuous E₂ with either continuous P₄ or discontinuous P₄. These regimens of hormone exposure were chosen to model ongoing clinical trials of HT in postmenopausal women that involve continuous E₂ with discontinuous delivery of P₄ for 10–12 days per month (14, 15, 27–29). Control animals were treated with vehicle (cholesterol). Hormones were delivered to rats via sc implanted SILASTIC brand capsules (1.57 mm inner diameter, 3.18 mm outer diameter) containing vehicle (cholesterol, 1 × 1 cm), 1.2 mg E₂ (15% E₂ in cholesterol, 1 × 1 cm), and/or 110 mg crystalline P₄ (3 × 3 cm) according to previously described methods (4, 30). Prior to implantation, capsules were primed overnight in sterile PBS to prevent an initial spike in hormone release. Cholesterol and hormones were purchased from Sigma-Aldrich.

Figure 2.

Treatment schedule. Hormone treatment began at day 0, 7 days after Sham OVX or OVX surgery, with continuous (Cont) capsule insertion. Discontinuous P₄ capsules (Discont) were inserted at day 20 and day 50 of the hormone treatment. The ECL lesion also occurred on day 50. On day 30, continuous capsules were replaced and discontinuous capsules were removed. Tissues were collected 64 days after the initiation of hormone treatment.

For the continuous hormone regimen groups, capsules were replaced on day 30. For discontinuous groups, P₄ SILASTIC brand capsules were inserted on day 20, removed on day 30 (at the time of E₂ insertion), and inserted again on day 50 (at the time of ECL surgery). SILASTIC brand capsules were inserted under general anesthesia (5% isoflurane for induction, 2%-3% isoflurane for maintenance). Mannino et al (30) previously characterized the pharmacokinetic profiles of SILASTIC brand capsule delivery of E₂ and P₄, reporting that the doses used in the current study resulted in plasma E₂ levels of approximately 70 pg/mL, comparable with peak levels observed at proestrus (31–33), and plasma P₄ levels of approximately 18 ng/mL, comparable with midrange levels observed during late proestrus (31–33). Previous studies have also demonstrated that the SILASTIC brand capsule method of delivery provides continuous delivery of a constant level of hormone for a period of months (30, 34, 35). Rats were divided into the following treatment groups: Sham + vehicle (Sham, n = 7); OVX + vehicle (OVX, n = 10); OVX + E₂ (O+E, n = 7); OVX + continuous P₄ (O+Pcont; n = 7); OVX + discontinuous P₄ (O+P, n = 7); OVX + E₂ + continuous P₄ (O+E+Pcont, n = 7); or OVX +E₂ + discontinuous P₄ (O+E+P, n = 7).

Animals were maintained at the University of Southern California vivarium facilities with ad libitum access to standard rat chow (Lab Diet; PMI 5001) and water. All experimentation was approved by the University of Southern California Animal Resources and Institutional Care and Use Committee and carried out in accordance with National Institutes of Health guidelines.

Entorhinal cortex lesion

On day 50 of the hormone treatment period, animals were administered a unilateral stereotaxic knife-cut lesion of the perforant path (ECL), resulting in partial deafferentation of the hippocampal dentate gyrus and retrograde degeneration of axotomized layer II neurons in the entorhinal cortex (Figure 2). Because both E₂ and P₄ have established neuroprotective actions, the lesion was administered on day 50 of the hormone treatment to ensure all P₄-treated groups would have P₄ present at the time of lesion. Tissues were collected 14 days after lesion EC cell loss, and neurite sprouting in the dentate gyrus has been previous characterized at this time point (13, 22, 24). Rats were anesthetized (70 mg/kg ketamine per 5 mg/kg xylazine, ip) and then placed in a stereotaxic frame with their heads in a level position and the incisor bar at −3.3 mm. The retracted wire knife (Scouten wire knife; Kopf, Tujunga) was positioned perpendicular to the midline (270°) and then inserted 0.5 mm anterior to lambda and 5.5 mm lateral to the sagittal suture such that the lesion was always in the left hemisphere (36). The knife was lowered 1 mm below the dura and rotated 30° clockwise (to 300° position), and then the knife was extended and a vertical cut was made by lowering the knife 2 mm. The knife was then retracted and rotated counterclockwise 165° (to the 135° position) before extending the knife again and making a vertical cut by raising the knife 2 mm. Lesion efficacy was confirmed by histological examination and differences within each animal between outcome measures on the ipsilateral vs contralateral brain regions. After removal of the knife, forceps were used to push the inner edges of skin at the incision site together and three to five wound clips were used along the length of the incision to fully close the wound site. Betadine was applied to the wound site and buprenorphrine (0.02 mg/kg, im) was administered postoperatively for analgesia. Fourteen days after the lesion, rats were transcardially perfused with ice-cold PBS followed by 4% paraformaldehyde while under general anesthesia (70 mg/kg ketamine per 5 mg/kg xylazine). The uteri were dissected to 1 cm length and weighed. The brains were collected and postfixed for 24 hours in 4% paraformaldehyde and then stored in PBS with sodium azide at 4°C.

Assessment of neuron viability

The brain was sectioned in the horizontal plane at 50 μm using a vibratome. For the assessment of neuron viability, every sixth section was immunostained with the neuron-specific antibody NeuN (1:250; Chemicon) using the ABC Elite immunohistochemistry kit, as previously described (4, 37). Control experiments in which the primary antibody was omitted revealed the expected absence of specific staining.

The numbers of NeuN immunoreactive cells in layer II of the entorhinal cortex were estimated using an unbiased optical dissector method as previously described (37). Every sixth section was collected within a 2-mm dorsal-ventral aspect of the EC surrounding the lesion. An Olympus BX50 microscope with a microcator and motorized stage controlled by CAST-grid software (Olympusenmark) was used to obtain unbiased, randomly orientated counting frames (32 × 32 μm with X-Y steps of 75 μm × 75 μm) throughout the layer II EC region. Within each counting frame, immunolabeled cells that exhibited intact nuclear morphology and that came into focus within the dissector height (z = −10 to −30 μm) were counted by an experimenter blinded to treatment group. The anatomical boundaries for quantification were modified from a previously established protocol (22), with one boundary being the point where EC layer II neurons form a distinct laminar organization emerging from the loosely organized parasubiculum and the lateral boundary stretching from the splenium of the corpus callosum to the apex of the cortical curvature between posterior and lateral cortices. The dorsal border was defined from the last point at which the knife lesion could be seen in the tissue section. Total neuron estimates were calculated according to the following equation: N = (ζQ⁻) (t/h) (1/asf) (1/ssf); where Q⁻ is the total number of neurons counted, t is section thickness (50 μm), h is dissector height (20 μm), asf is the area sampling fraction, and ssf is the section sampling fraction (1/15). NeuN-labeled cells in EC layer II were counted in both the ipsilateral (lesioned) and contralateral hemispheres (unlesioned), which allowed the calculation in each animal of the percentage of cell survival (ratio of cell numbers in lesioned relative to unlesioned hemisphere).

To qualitatively verify NeuN findings, a separate set of tissue sections from all animals were thionin, a general stain for DNA and Nissl substance. Sections were mounted onto gelatin-subbed slides and air dried overnight and then stained with 0.1% thionin, dehydrated with ethanol solutions, and coverslipped with Krystalon (EMD Chemicals).

Assessment of neurite sprouting

To assess compensatory neurite sprouting in the DG, three 50-μm sections were selected from each animal sample and stained with the Holmes fiber method using a previously described method (38). Briefly, sections were mounted on gelatin-subbed slides and air dried overnight. Rehydrated sections were bathed in 1% silver nitrate solution and incubated overnight at 37°C in impregnation solution (0.2 M boric acid buffer, 0.1 M borax buffer, 1% silver nitrate, and 10% pyridine). The following day, slides were bathed in reducing solution (0.09 M hydroquinone and 1.3 M sodium sulfite) followed by washes in 0.2% gold chloride, 2% oxalic acid solution, and 5% sodium thiosulfate. Finally, stained sections were dehydrated in a series of graded ethanol solutions, cleared with xylene, and coverslipped with permanent mounting medium.

The width of Holmes-positive neurite growth extending from the DG granule cell layer was measured. Slides were viewed using a Leica DMLB microscope (McBain Instruments) coupled to a MTI 3CCD camera (Dage-MTI), and images were captured using IPLabAlias software. The method of quantification was modified from a previously described protocol (13). In brief, Image J (National Institutes of Health, Bethesda, Maryland) was used to make three measurements from each image (five images per section, three sections per animal), representing the apparent longest, shortest and an average width of fiber outgrowth, as detected by Holmes fiber staining. These three measurements of neurite outgrowth width were averaged and used to calculate a percentage of total width. The total width was measured from the granule cell layer to the hippocampal fissure.

Statistical analyses

Data were analyzed using the Statistical Package for Social Sciences (version 11.5; SPSS Inc). All data are presented as mean ± SEM. One outlier from the O+P group was excluded from analyses. One-way ANOVA was performed to determine the effect of hormone status on percent neuronal loss and fiber staining. Homoscedasticity was assessed with Levene's test (neuronal survival, P = .487; neurite outgrowth, P < .001). Tukey-Kramer post hoc tests were used for between group comparisons of parametric data (neuronal survival), Games-Howell post hoc tests were used for between group comparisons of nonparametric data (neurite outgrowth). Significance was set at a threshold of P < .05.

Results

In this study, we examined the effects of OVX-induced ovarian hormone depletion and various E₂ and P₄ hormone treatment regimens on neuron survival in EC layer II and compensatory neuronal sprouting in the dentate gyrus after ECL. As a bioassay of estrogenic effects, body and uterine weights were assessed, with a significant uterotrophic effect of E₂ but not P₄ observed, confirming the efficacy of the E₂ SILASTIC brand capsules (Supplemental Table 1).

Neuron viability

To investigate the effects of treatment conditions ECL-induced neuron loss in EC layer II, we compared stereological estimates of NeuN immunoreactive cells in lesioned (ipsilateral) relative to unlesioned (contralateral) hemisphere. In the absence of hormone manipulation (Sham), ECL resulted in a modest but significant reduction in the total number of EC layer II neurons in the lesioned (65 937 neurons ± 9290) compared with unlesioned hemisphere (76 172 neurons ± 9228; paired t test: t = 2.992, P = .024; Figure 1, B and C). Total neuron numbers in the unlesioned hemisphere were not affected by hormone manipulation (F = 0.34, P = .91; Figure 3A). Neuronal survival in the lesioned hemisphere was significantly affected by sex hormone status [F_{(6, 50)} = 2.87, P < .02; Figure 3, B–I]. Mean neuronal survival in vehicle-treated OVX rats was not significantly different from Sham (P = .11), O+E (P = .28), O+Pcont (P = .67), O+P(P = .79), or O+E+Pcont groups (P = .98). E₂ combined with discontinuous P₄ (O+E+P) was the only hormone treatment to significantly increase the EC layer II neuronal survival in OVX rats (P < .02 vs OVX). The relative differences in neuron survival across the groups were confirmed with thionin staining (Supplemental Figure 1).

Figure 3.

E₂ combined with discontinuous P₄ promotes neuron viability after the ECL. A, Total neuron numbers in the EC layer II were unaffected by hormone manipulations in the contralateral hemisphere (unlesioned). B, Percentage neuron survival in the lesioned relative to the unlesioned hemisphere was significantly altered across treatment groups. E₂ supplementation combination with discontinuous P₄ (O+E+P) significantly reduced neuronal loss compared with OVX rats administered vehicle (OVX). C–I, Representative photomicrographs show NeuN immunoreactivity in the EC layer II after the ECL in Sham-OVX (C, Sham) and OVX rats administered vehicle (D, OVX), continuous E₂ (E, O+E), discontinuous P₄ (F, O+Pcont), discontinuous P₄ (G, O+P), E₂ combined with continuous P₄ (H, O+E+Pcont), and E₂ combined with discontinuous P₄ (I, O+E+P). Data are presented as means ± SEM. *, P < .02 compared with OVX group. Scale bar, 100 μm.

Neurite outgrowth

In the unlesioned hemisphere, Holmes fiber staining revealed neurite networks that extended from the DG granule cell layer throughout the entire molecular layer to the hippocampal fissure (Figure 1D). In the lesioned hemisphere, reduced fiber staining was observed in the molecular layer of the DG, with stained neurite networks terminating in the inner molecular layer (Figure 1E). To determine the effect of P₄ delivery schedule on adaptive responses in the DG following ECL, we compared the effects of E₂ alone or in combination with either continuous or discontinuous P₄ on compensatory neuronal sprouting. Staining of neurite outgrowth was significantly affected by sex hormone status [F_{(6, 50)} = 16.511, P < .001; Figure 4]. Compared with the Sham group, OVX reduced fiber staining by more than 40% (P < .05). E₂ combined with discontinuous P₄ was the only hormone treatment to significantly increase neurite sprouting in OVX rats (O+E+P, P < .05), rescuing neurite sprouting to sham levels (P = 1 vs Sham). Fiber outgrowth in rats treated with E₂ combined with discontinuous P₄ (O+E+P) was significantly increased compared with all other hormone treatments (P < .05).

Figure 4.

E₂ and P₄ promote neurite sprouting after the ECL. A, Quantification of neurite fiber staining in the DG of the hippocampus after the ECL shows significant differences across treatment groups. B–H, Representative photomicrographs of Holmes fiber staining in the DG after the ECL in gonadally intact (B, Sham) and OVX rats administered vehicle (C, OVX), E₂ alone (D, O+E), continuous P₄ alone (E, O+Pcont), cyclic P₄ alone (F, O+P), E₂ combined with continuous P₄ (G, O+E+Pcont), and E₂ combined with cyclic P₄ (H, O+E+P). Data are presented as means ± SEM. gcl, granule cell layer; hf, hippocampal fissure. *, P < .05 relative to OVX group; †, P < .05 relative to O+E+P group. Scale bar, 50 μm.

Discussion

In this study, we demonstrate that varying P₄ delivery schedule in combined E₂ and P₄ treatments alters degenerative and regenerative outcomes in female rats after unilateral ECL, a model of AD-like EC layer II cell loss and DG deafferentation. Our findings suggest that that the combined efficacy of E₂ and P₄ on neuroprotective and regenerative outcomes is dependent on the administration regimen. Importantly, the discontinuous-combined E₂+P₄ regimen was the only treatment condition to improve neuronal survival and neurite outgrowth in the OVX rats. Prior work demonstrated that short-term interactions between E₂ and P₄ are important in the regulation of a variety of neural sex steroid actions including sexual receptivity (39), gonadotropin secretion (40), and synaptic plasticity (41). The current findings extend a growing literature that indicates neural effects of long-term E₂ and P₄ exposure also depend on the treatment regimen, the understanding of which has important implications for HT in postmenopausal women.

Diverse protective actions of both E₂ and P₄ have been described in several models of neural injury (42–49). In the current study, neither E₂ nor P₄ significantly promoted EC layer II neuronal survival or neurite outgrowth when administered independently. Although the absence of E₂ neuroprotection is unexpected, it may be more reflective of a modest lesion than lack of hormone benefit. Note that OVX-induced hormone depletion showed only a nonsignificant trend of reduced neuron survival, which showed a nonsignificant trend of reversal by treatment with E₂ alone. Notably, E₂ combined with discontinuous (but not continuous) P₄ significantly increased both neuronal survival and neurite outgrowth. One potential limitation of these data is that neuronal viability was quantitatively assessed using the postmitotic neuronal specific nuclear protein NeuN, which can be down-regulated in surviving neurons under stress (50, 51) and thus may underestimate neuronal survival after injury. To address this possibility, NeuN findings were qualitatively confirmed with thionin, a general cell stain that labels DNA and Nissl bodies.

In general agreement with the current findings, we have previously observed that continuous P₄ attenuated E₂-mediated neuroprotection after kainate lesion (3, 4). The mechanism by which continuous P₄ inhibits E₂/P₄-mediated neuroprotection remains unclear, although previous studies have demonstrated that P₄ attenuates E₂-induced elevations in key neurotrophins including brain-derived neurotrophic factor (BDNF) (5, 6). For example, in cultured hippocampal slices, E₂ protected against N-methyl-D-aspartate toxicity by the up-regulation of BDNF expression and activation of the BDNF receptor tropomyosin receptor kinase B, whereas P₄ antagonized these protective actions of E₂ (5). BDNF has been shown to reduce EC layer II cell loss after ECL (52), suggesting that interactions between E₂ and P₄ on BDNF expression may contribute to the observed effects on neuron survival.

Ovariectomy significantly reduced compensatory sprouting after ECL-induced deafferentation in the DG, consistent with previous findings (13, 53, 54). Prior work suggests that the ovarian hormones, E₂ and P₄ may promote neurite outgrowth through several mechanisms, including E₂ mediated effects on apolipoprotein E (55, 56), P₄ mediated promotion of myelination (57–59), and P₄ induced proliferation of neural progenitor cells derived from the DG (60, 61). Newborn neurons have increased plasticity compared with mature neurons after ECL, innervating the inner molecular layer at a higher density after ECL (62). However, in the current study, the effect of OVX was not reversed by E₂ or P₄ administered alone or combined in a continuous treatment regimen. These findings differ from observations by Wong et al (13), who found increased synaptic sprouting in the DG of continuous E₂- but not P₄-treated OVX rats after ECL lesion, with E₂-mediated neurite sprouting attenuated by combination with continuous P₄. These discrepancies may be the consequence of key differences in hormone doses and delivery regimens. First, Wong et al administered E₂ and P₄ using a pellet delivery system that has been reported to release a hormone surge in the first week of administration that is 5- to 15-fold higher than more physiological levels observed subsequent weeks (63). Second, the use of two consecutive cycles of hormone treatment in the current study prior to the lesion surgery may have altered the efficacy of the neurite-promoting actions of E₂.

E₂ combined with discontinuous P₄ was the most beneficial regimen for both EC layer II neuron survival and DG neurite sprouting. However, because discontinuous P₄ groups underwent one additional capsule insertion under isoflurane anesthesia (d 20, 1 mo prior to lesion), we cannot exclude that the increased efficacy of the E₂ combined with discontinuous P₄ treatment could potentially be attributed to indirect effects of the additional exposure to anesthesia. For example, isoflurane anesthesia has been shown to stimulate adrenal corticosteroid production and can modulate expression of BDNF (64, 65), although this effect may be minimal with brief exposure, as is required for capsule insertion (66). Furthermore, isoflurane has been demonstrated to elicit diverse effects on the hippocampus and cortex, ranging from increased susceptibility to neuronal degeneration (67, 68) to increased resistance to neurotoxic insult (69, 70). Despite this limitation, the findings are consistent with our previous findings of qualitative differences in the neural effects of discontinuous vs continuous P₄ regimens. For example, E₂ combined with discontinuous but not continuous P₄ reduces β-amyloid accumulation in 3xTg-AD mice (12, 14) and in wild-type rats (27). Others have reported that E₂ combined with discontinuous but not continuous P₄ improves cholinergic function in OVX female rats (72). Recent evidence suggests that observed qualitative differences in functional outcomes across P₄ regimens are related to underlying changes in gene expression (15).

Although the mechanisms underlying interactions between P₄ and E₂ remain incompletely defined, in vitro and ex vivo studies suggest P₄ may modulate estrogen receptor (ER) expression. In cultured hippocampal slices, E₂ treatment up-regulated ERβ expression, whereas priming with the P₄-blocked E₂ induced the up-regulation of ERβ and inhibited E₂-mediated neuroprotection (73). In primary hippocampal neuron cultures, P₄ decreased the expression of both ERα and ERβ as well as ER-dependent transcriptional activity (74). Likewise, the progesterone receptor membrane component-1 (Pgrmc1) and the classical P₄ receptor (Pgr) can be regulated by both E₂ and P₄ (15, 75). For example, Bali et al (75) reported that mRNA and protein levels of Pgrmc1 were increased in the DG by independent and combined treatments of E₂ and P₄. Zhao et al (15) demonstrated that the effects of E₂ and P₄ on Pgr and Pgrmc1 mRNA expression are dependent upon the P₄ treatment regimen, such that continuous but not discontinuous P₄ (alone or in combination with E₂), down-regulated Pgr expression.

Our results indicate that differences in the P₄ administration regimen can markedly alter the neuroprotective and regenerative outcomes in female rats, findings that may be predictive of the neural efficacy of HT regimens in postmenopausal women. Consistent with this possibility, Tierney et al (76) reported E₂ combined with intermittent progestin (norethindrone) prevented declines in delayed verbal recall in women with normal memory function. The recent clinical trials Kronos Early Estrogen Prevention Study (29) and Early vs Late Intervention Trial with Estradiol (77) should provide further insight into the potential cognitive benefits of HT with discontinuous P₄. Continued preclinical and clinical investigation of this issue is needed to optimize the therapeutic potential of HT against age-related cognitive decline and neurodegenerative disorders including AD.