Hormone Therapy and Effects on Sporadic Alzheimer’s Disease in Postmenopausal Women: Importance of Nomenclature

DEFINITION OF ALZHEIMER’S DISEASE

Even though a preponderance of pre-emptive measures have been forwarded to delay cognitive decline and Alzheimer’s disease (AD) in our aging population, to date there is no effective medical intervention or defensive measure in place to prevent or cure AD in our society.

AD is a debilitating and insidious disease which results in the sequential destruction of neurons and synaptic connections, development of dementia, brain atrophy, and finally death [1]. Microscopic pathological characteristics of AD include the presence of two anomalous structures: neuritic plaques and neurofibrillary tangles. Neuritic plaques are formed by the deposition of clumps of amyloid-β (Aβ) proteins which build up in neural interstitial spaces eventually preventing communication between nerve cells [2]. In addition, neuronal cell functioning is dependent on the structurally supportive, nutritive, and other essential transport actions of microtubules formed by the protein tau. In AD, it is thought that hyperphosphorylation of the tau proteins causes the breakdown of microtubules to form aggregates of paired helical filaments called neurofibrillary tangles [3].

HORMONE STUDIES

Numerous observational studies have suggested that hormone therapy (HT) might protect postmenopausal women against cognitive decline and AD. However, because of the significant disparity between results, especially those between observational and randomized controlled trials (RCT), this postulate remains unproven. A significant contributing factor to these inconsistencies is the loose use of the generic definitions of estrogens and progestogens [4, 5]. This is highlighted by the generalized terminological use of HT, which is often used to encompass significantly disparate hormonal formulations without clear delineation. For example, there is a significant difference between non-endogenous estrogens, e.g., conjugated equine estrogens (CEE) and endogenously identical (bioidentical) 17 beta-oestradiol (E2) [6, 7]. Similarly, progestogens (which encompasses progesterone (P4) and progestins) are also often not clearly demarked despite bioidentical P4 and progestins, e.g., medroxyprogesterone acetate (MPA) having significantly disparate molecular structure, pharmacological, and physiological actions [8, 9]. There is a significant difference between hormones that exhibit exact receptor binding properties (e.g., E2 and P4) and the resultant molecular and physiological consequences of those actions compared to those that do not (e.g., CEE and MPA). The substantial and ever-growing body of evidence continues to reinforce that non-endogenous estrogens (e.g., CEE) and progestins (e.g., MPA) do not have the same beneficial experimental, prospective, and clinical outcomes as bioidentical E2 and P4 on brain and other physiological functioning [10–15]. The differential actions between progesterone (P4) and medroxyprogesterone acetate (MPA) are clearly delineated in Table 1 [16].

Additionally, the noteworthy beneficial effects of primarily unopposed (used alone) transdermal E2 are reinforced in clinical and prospective studies with significant cognitive improvement in various domains, including verbal memory, working memory, attention, and visuo-spatial abilities being noted in post-hysterectomized and bilaterally oophorectomized women (10 mg/day (d) intramuscular E2 valerate) [17]; in elderly, healthy, postmenopausal women, E2 (100μg/d/2 weeks, within treatment, i.e., E2 levels >29 pg/mL, age 69±1.4 years (y) [18], E2 (100μg/d/3 weeks, age 55–75 y) [19], E2 (100μg/d/8 weeks, age 56–84 y) [20]; in postmenopausal women at risk for AD (E2^* versus CEE±P4^*) [21, 22]; postmenopausal women with mild cognitive impairment E2 (0.5 mg/d–2 mg/d/24 months)+opposed (oral P4 100 mg/d), age 57–82 y [23] and in postmenopausal women with AD, E2 (50μg/d/13 weeks, age 66–89 y) [24], E2 (100μg/d/16 weeks, age 61–90 y) [25], E2 (50μg/100μg/d)±opposed MPA (2.5 mg/d)/12 months, age 55–85 y) [26]. (^*mode of administration not specified).

In contrast, no cognitive improvements were noted in a number of RCTs using unopposed CEE versus placebo. A 16-week RCT of 42 postmenopausal women, with mild to moderate AD, demonstrated no cognitive improvement after treatment with 1.25 mg/d unopposed CEE versus placebo [27]. Additionally, a 12-week trial of 50 postmenopausal women with AD treated with unopposed 1.25 mg/d CEE demonstrated no effect on cognitive performance, dementia severity, mood, behavior or cerebral perfusion compared to placebo [28]. This data is consistent with the findings of the Women’s Health Initiative Memory Study (WHIMS) [29–32]. In WHIMS [29–32], two arms of investigation determined the effects of HT on postmenopausal women, 65 years and older. One arm (4,532 women) received oral CEE (o-CEE) in combination with MPA or a placebo. The other arm comprised hysterectomized women (n = 2,947) who received unopposed CEE or placebo. The results clearly demonstrated that postmenopausal women, who received CEE in combination with MPA or CEE alone, were not protected against dementia or cognitive decline but instead were at increased risk for probable dementia and cognitive impairment [33].

CEE hormonal formulations, derived from pregnant mares’ urine, used in the Women’s Health Initiative (WHI) [34, 35] and WHIMS [29–32], is a complex formulation of at least ten identified estrogens with other estrogens and components yet to be fully elucidated [7, 36]. Identified estrogens in CEE include estrone sulphate and the equine specific, equilin and equilenin [7, 36]. Hormonally, it has been determined that estrone (E1) is the most predominant estrogen in postmenopausal women [37]. It is therefore suggested that this hormone is neither disease preventative nor protective. Studies demonstrate that during menopause there is a marked decrease in the levels of the primary endogenous reproductive hormones, E2 and P4, and as a consequence there is a shift in the ratio of E2 and E1, with E1 becoming the primary estrogen in postmenopausal women [37]. The change in ratio between E1 and E2 significantly affects receptor binding processes and as a consequence molecular and physiological processes are also notably impacted. For example, the hippocampus contains the classical α and β estrogen receptors (ER) [38–41]. E2 is the most potent estrogen and binds to the classical estrogen receptor with a significantly greater affinity than E1. That is, E1 has a 40-fold lower affinity for ERα and approximately a 60-fold lower affinity for ERβ than E2 [42, 43]. It has also been demonstrated that E2, not E1, increases survival and activation of new neurons in the hippocampus in response to spatial memory in adult female rats [43].

In summary, the significant molecular and subsequent physiological and cognitive outcome differences between E2 and CEE emphasize that the actions of these HT formulations are not homologous but require individual assessment.

IMPACT OF THE MODE OF HORMONAL ADMINISTRATION

The mode of hormonal administration (transdermal/percutaneous versus oral) has also been shown to significantly impact physiological processes and safety and as a consequence may also be a significant contributor to the differential cognitive effects noted between studies. Studies have consistently demonstrated that oral E2 does not confer the same beneficial physiological effects as transdermal E2. First pass liver metabolism, as a consequence of oral administration, results in increases in metabolic risk markers including increases in triglycerides, decreases in low density lipoprotein (LDL) particle size, higher levels of C-reactive protein (CRP), and activation of coagulation [11]. Substantial increases in plasma E1, sex steroid binding protein (SBP), renin substrate, very low density lipoprotein (VLDL), and significant decreases in anti-thrombin activity (AT) have been noted after treatment with oral micronized E2 and oral E2 valerate, but no change after treatment with E2 percutaneous administration [44]. Oral, but not transdermal E2, increased platelet reactivity in a subset of postmenopausal women [45] and oral, but not transdermal E2, was associated with an increased risk of venous thromboembolism (VTE) in postmenopausal women [46]. Oral E2 (1 mg), not transdermal E2 (50μg), induced activated protein C (APC) resistance and activated blood coagulation [47] and compared to transdermal testosterone (12.5 mg/d) and intra-vaginal progesterone (100 mg/d) which showed no effect, transdermal E2 (0.1 mg/d) treatment demonstrated significant reduction in very low density lipoprotein-triglyceride (VLDL-TG) concentration due to accelerated VLDL-TG plasma clearance (25.1±2.5 versus 17.4±2.7 mL/min; p < 0.01) [48]. Contrary to the beneficial effects elicited on various cognitive domains using primarily unopposed transdermal E2 (Section: Hormone studies), compared to placebo, no cognitive benefit has been noted using unopposed oral E2 (0.5–2 mg/d) [49] or unopposed oral E2 (1 mg/d) or 1 mg/d oral E2 opposed by 5 mg/d MPA (12d per y) [50] or unopposed oral E2 (2 mg/d) or 2 mg/d E2 opposed by 100 mg/d oral P4 [51]. The Early versus Late Intervention Trial (ELITE) comprising 567 women within 6 (early) or 10 + y (late) postmenopause demonstrated no cognitive benefit or harm after treatment with oral E2 (1 mg/d)±cyclical intra-vaginal P4 (45 mg/d/non hysterectomized) versus placebo for a period of 57 months [52]. Attaining physiological reproductive levels of E2 is also important in producing favorable cognitive outcomes. For example, no benefit or cognitive harm was noted in healthy postmenopausal women (aged 60–80 y) after 2 years administration of a sub-physiological dose of unopposed transdermal E2 (14μg/d) [53]. No cognitive benefit or harm was also noted after 3 y treatment of 57 postmenopausal women (average age: placebo 75±4 y, E2 treated 76±6 y) with ultra-low dose E2 (25μg/d) with all non-hysterectomized women receiving oral P4 (100 mg/d/2-week period every 6 months) [54]. Furthermore, due to the marked reduction in hormone levels postmenopause, increases in insulin and cortisol levels have also been noted. These disruptions in insulin functioning have been associated with the development of diseases including AD, cardiovascular disease, and type 2 diabetes [55, 56]. Studies have demonstrated that the elevated levels of insulin and cortisol can be normalized by treatment with transdermal E2 [57] but remain or increase in elevation after treatment with oral E2 [58]. It has been shown that E2 regulates insulin degrading enzyme (IDE) hippocampal expression via an ERβ/P13-K pathway in 12-month-old 3xTg-AD ovariectomized (Ovx) mice [59]. IDE is not only important for degrading insulin but also for the clearance of Aβ from the brain [60].

WINDOW OF OPPORTUNITY HYPOTHESES

Following the results of the WHI [34, 35] and WHIMS [29–32], the critical window and healthy cell bias [85] hypotheses were forwarded to try and reconcile the disparate results between these major studies and those that demonstrated positive effects after hormonal administration. In reference to the WHIMS findings, it was surmised that cognitive and neurological benefit would be afforded if hormonal therapy was administered to women during the menopause transition or in early postmenopause and adverse if treated later in the postmenopausal process (65 + y). These postulates have persisted even though a number of prospective and clinical studies, pre and post WHI and WHIMS, have demonstrated significant cognitive improvement in various domains including verbal memory, working memory, attention and visuo-spatial abilities, including women 65+, after treatment with primarily unopposed transdermal E2 (Section: Hormone studies). Instead of recognizing that the reasons for WHI and WHIMS results were most likely due to the women being treated with hormones that were not identical to their own endogenous hormones and having different receptor binding affinities and functionalities; validation of critical window and healthy cell bias hypotheses came from various sources. These included, investigation of the effects of various hormonal formulations at the time of the menopause transition or in early postmenopause, e.g., the Knonos Early Estrogen Prevention Study (KEEPS) [86, 87], to see if it would yield more favorable results and some animal experiments using bilaterally Ovx animals, primarily rodents. In respect to the latter, numerous rodent studies demonstrated that if animals were treated with unopposed E2 immediately post or shortly after surgical removal of the ovaries, then normal investigated processes were re-established. However, if a considerable amount of time had elapsed post-surgery, functionally was either impaired or not restored [69, 88–93]. The reasons for these results are most likely multi-factorial with one of the major factors being that bilateral Ovx of these animals would result in an immediate and complete cessation of ovarian function and the loss of essential steroidal hormones and substrates. If these are not replaced exogenously, immediately post-surgery, then over time this would have serious long-term physiological consequences, not only for the brain, but also other major organs that rely on these ovarian steroids for functionality. In postmenopausal women, even after the proposed postmenopausal two-year nadir for estrogen has been reached, the ovaries are still essential for the production of steroidal hormones and precursors [94–96]. Hence, in postmenopausal women, a cessation of follicular not ovarian function is a more appropriate evaluation. Therefore, even though animal studies have provided fundamental information regarding the beneficial effects of E2 and P4 on brain function and dysfunction (Section: Experimental actions of E2 and P4), the results from Ovx animals cannot be extrapolated to simulate what happens under natural conditions to postmenopausal women who have not undergone bilateral oophorectomy, nor as an argument to support the window of opportunity hypotheses.

What is significant to note from human studies and gives a strong insight of the immense importance of hormones, specifically E2 and P4 on brain functioning, was the physiological consequences of women who had undergone oophorectomy prior to natural menopause [97]. Studies by Rocca et al. [98] have demonstrated that unilateral and bilateral oophorectomy before the onset of menopause was associated with an increased risk of cognitive impairment or dementia compared to women that had never undergone an oophorectomy. For associated studies, refer to [99–103]. Shuster et al. [104] also emphasized that premature and early menopause, whether spontaneous or induced, have been associated with long term health risks such as neurologic disease, mood disorders, cardiovascular disease, and premature death. These findings are reinforced in the studies by Sherwin [105] and Phillips & Sherwin [17]. An RCT conducted by Phillips and Sherwin [17], determining the effects of 10 mg intra-muscular E2 valerate on various cognitive domains (assessed pre- and post-surgery) in total hysterectomized and bilaterally oophorectomized women from benign causes, demonstrated significant differences between treated versus placebo in immediate and delayed recall of paired-associates. Those treated with E2 demonstrated no significant changes in these scores pre- and post-surgery, which showed that E2 status and its specific cognitive influence, was reinstated post-surgery. This was contrary to the placebo group where those specific scores were significantly decreased post-surgery compared to pre-surgery. It was also noted that compared to placebo, treatment with E2 did not impact immediate or delayed recall of visual material, delayed recall of paragraphs or digit span scores. Additionally, an earlier cross-over designed RCT by Sherwin [105] on hysterectomized and bilaterally oophorectomized women (from benign causes) using either IM E2 valerate (10 mg/d), androgen or combined estrogen-androgen, or placebo demonstrated that women treated with all three hormonal replacements maintained their performance post-surgery on tests of digit span, clerical speed and accuracy, paragraphs of words recalled, and abstract reasoning. In contrast, significant decreases in these memory scores were noted in those treated with placebo post-surgery (p<0.01). The above results therefore give an important insight into the physiological impact of bilateral oophorectomy in reproductive women as a result of immediate cessation of ovarian function and the induction of abrupt menopause. These studies also highlight the importance of the protective hormones E2 and P4 (in combination with testosterone) in cognitive functioning and how the abrupt loss, which has been associated with a high risk of developing AD, may give a potentiated insight into the dementia process.

The four-year multicenter, randomized, double blind, placebo controlled, KEEPS, comprising 727 healthy, menopause transition and early postmenopausal women (6 to 36 months post last menses) has provided important information on the effect of bioidentical transdermal E2 (t-E2) (50μg/d) via a Climara^® patch versus oral CEE (o-CEE) (Premarin 0.45 mg/d) and cyclical oral micronized P4 (Prometrium^®) (200 mg/d administered 12 days each month to estrogen treated women) [86, 87]. Various physiological domains (including atherosclerosis progression over time) in women aged 42–58 y were investigated. Detailed assessments of cognitive function, depressive symptoms, effects of mood and sexual function were investigated in the cognitive and affective arm of the study (KEEPS-Cog). Six hundred and ninety-three women who at baseline were free of depression, dementia, and memory deficits underwent comprehensive battery tests measured at baseline and at 18, 36, and 48 months. In the physiological arm of the trial, results demonstrated those treated with o-CEE exhibited increases in high density lipoprotein (HDL) and decreases in low density lipoprotein (LDL) levels, but also increases in triglyceride and C-reactive protein (CRP) levels. In contrast, those treated with t-E2 showed improved glucose levels and insulin sensitivity and no adverse effects on other investigated biomarkers [106–108]. In the KEEPS-Cog arm of the trial, no adverse effect on domain-specific or general measures of cognition were noted. However, as cited earlier, even though no cognitive deficits have been noted to date [109], MRI analysis of a small subset of women (n = 95) noted rates of ventricular volumes increased in o-CEE treated women versus placebo (p = 0.01) [110]. Additionally, three years after stopping randomized treatment, lower levels of Aβ (particularly in APOE ɛ4 carriers) were noted in t-E2 treated women compared to o-CEE treated women and placebo [111]. It was also noted that greater preservation of the dorso-lateral prefrontal cortical volume correlated with lower global PET PiB uptake in the t-E2 group only [112]. Reduction in total brain volume, comparatively assessed by MRI, has also been noted in a subset of 1,402 women (WHIMS MRI study) at 2.4 years post the WHI trial (CEE±MPA versus placebo) and repeated in 699 women 4.7 years later [113]. Compared to placebo those with type 2 diabetes had the most marked decreases in brain volume (decrement of –18.6 mL) while those without diabetes had decrements of –0.4 mL. For associated papers, refer to [114, 115].

An issue with the KEEPS [86, 87] methodology that requires consideration was the use of an oral rather than an intra-vaginal FDA approved P4 formulation. The oral route requires the hormone to undergo first pass intestinal wall and hepatic metabolism [116, 117]. Under natural conditions, which the intra-vaginal route more closely approximates, the endogenous ovarian hormones enter the circulation via the pelvic plexus of veins, are carried to the heart and lungs then circulated around the body to the liver where they are metabolized and detoxified [118].

Progesterone (P4) has been shown to have substantive neuroprotective properties [72, 120]. The 5α-isomer P4 metabolite allopregnanolone has also been shown to exhibit neuroprotective actions [121]. Studies have also demonstrated that the mode of administration of P4, whether oral versus intra-vaginal, can significantly impact metabolic pathways, serum concentrations, and resultant metabolites. Plasma levels of P4 increase rapidly with both vaginal and oral administration; however, plasma levels were significantly higher with vaginal administration compared with oral administration and levels were sustained for a significantly longer period [116]. Additionally, after vaginal administration, plasma deoxycorticosterone (DOC), a potent mineralocorticoid, did not significantly differ from mean baseline levels. In contrast, after oral P4 administration DOC levels were noted to be significantly higher than mean baseline levels [117].

5α-dihydroprogesterone metabolizes to allopregnanolone or iso-allopregnanolone, whereas 5β-dihydroprogesterone metabolizes to pregnanolone or iso-pregnanolone [119]. Progesterone binds to the classical progesterone receptor (PR), while allopregnanolone is a potent ligand of the GABA-A receptor. In contrast, iso-pregnanolone does not bind directly to the GABA-A receptor but it antagonizes the effect of allopregnanolone on the GABA-A receptor [122]. It has been demonstrated that intra-vaginal administration primarily promotes metabolism via the α metabolic pathway while oral P4 administration promotes metabolism through both α and β pathways [117]. The differential metabolic and metabolite pathways between oral versus intra-vaginal P4 administration therefore reinforce that the intra-vaginal methodological choice would promote more significant beneficial effects on cognition compared to oral administration. In addition, unlike transdermal routes of P4 administration which have shown inconsistent absorbency profiles and variable ability to suppress the proliferative effects of estrogen on the endometrium [123, 124], intra-vaginal P4 applications have demonstrated more favorable absorption profiles and suppressive uterine estrogenic proliferative actions [125–128].

EXPERIMENTAL ACTIONS OF E2 AND P4

The beneficial effects of E2 and P4 on neurotransmission, neuroconnectivity, and neuroprotective function on brain function and dysfunction, especially regarding AD, are substantive (Table 2).

As outlined in Table 2, the actions of E2 and P4 on brain functioning are significant. Therefore, when the levels of these two hormones become markedly depleted postmenopause, glial and neuronal processes which are reliant on the hormones for optimum functionality will be notably compromised. This has major implications for the development of AD.

It is therefore proposed that the restoration of lower range reproductive levels of transdermal E2 and cyclically administered intra-vaginal P4 in postmenopausal women exhibiting mild AD should result over time in the reestablishment of their hormonal status and protective physiological functionality.

CONCLUSION

Considering the significantly disparate receptor binding actions and the resultant molecular, pharmacological, and physiological consequences of those actions between bioidentical and non-bioidentical estrogen compounds and progesterone and various generations of progestins, it is no longer acceptable to group HT formulations together as a homogenous group. It should be made clear what type of compound/s were used in the research, mode of administration (e.g., transdermal versus oral versus intra-vaginal), dosage, opposed or unopposed, whether administered cyclically or continuously, what hormone levels were achieved, etc. This has significant implications when interpreting data, meta-analyses, observational studies, etc., relevant to AD and other disease processes including breast cancer, cardiovascular disease, etc. This will also allow differentiation between compounds that exhibit beneficial actions and those that do not and whether these effects are specific or generalized. For example, molecularly disparate progestins function is to downregulate the proliferative effects of estrogen on the uterus. However, what are their individual long-term effects on other hormone responsive tissues such as the brain, cardiovascular system, mammary glands, etc., and numerous metabolic processes? [11, 169–195].

The significant molecular and resultant physiological differences between HT formulations also highlight that the two hypotheses, the critical window [84] and healthy cell bias [85], which were forwarded to try and explain the detrimental effects of CEE and MPA used in the WHI and WHIMS cannot be extrapolated to encompass the actions of all estrogen and progestogen formulations on physiological functioning.

An area that requires further investigation, and has been raised as an important issue numerous times in solid argument statements and critical reviews, is that neurological sex differences are relevant and even though the end stage AD neurological pathologies between male and females may be similar, the processes involved in AD progression may differ significantly [196–217]. This may have relevance to our understanding of AD degenerative processes but also the development and utilization of biomarkers. It also raises a serious question why experimental hormonal data from male animals are being used to explain women’s hormonal processes in the area of AD, cardiovascular disease, etc., research. Data from male rodents therefore cannot be extrapolated to simulate female physiology either in female rodents or human females [218].

Decades of research have yet to yield the etiology of sporadic AD. Molecular research, fundamental and often of brilliance, is unable to differentiate between degenerative processes that are causal and those that are consequential. It is therefore important, perhaps in an interdisciplinary approach, to step back and focus on fundamental questions relevant to the processes of disease development. Firstly, it is of note that two-thirds of those exhibiting sporadic AD are postmenopausal women [198, 220]. As stated above, it is also proposed that the reproductive phase of a women’s life is protective against sporadic AD. When the levels of E2 and P4 are significantly disrupted during perimenopause and drop markedly postmenopause, this dimorphic protective status is no longer afforded [75–81].

In women’s postmenopausal health, bioidentical hormones have often been portrayed negatively despite there being numerous formulations that are PGA and FDA approved, e.g., E2 patches and intra-vaginal applications, etc., and P4 oral formulations and intra-vaginal applications, etc. Bioidentical hormones can also be obtained, via prescription, from compounding pharmacies. It is of note that research, using E2 and P4 formulations, published in high ranking, peer reviewed scientific journals, in the area of basic, experimental, clinical, AD, cardiovascular, and breast cancer research is substantive. As reiterated above, bioidentical E2 and P4 are molecularly identical to our endogenous hormones which, during the reproductive phase, have been protecting women since time immemorial.

Baseline assessment of parameters such as mammogram, uterine ultrasound, hormone levels, blood analyses, cardiovascular assessment, etc., prior to menopause hormonal administration would facilitate detection of anomalies and allow evaluation of these parameters at regular intervals post hormonal administration to ensure safety and wellbeing of treated women. To date there are no PGA or FDA approved P4 transdermal patches and therefore the importance of achieving P4 levels with prescribed compounded formulations that can negate the proliferative effects of E2 on the uterus, mammary glands, etc., cannot be overemphasized [128, 220].